Behav Ecol Sociobiol (1991) 29:235-247 Behavioral Ecology and Sociobiology © Springer-Verlag 1991

The morphology and behavior of dimorphic males in Perdita portMis (Hymenoptera: Andrenidae) Bryan N. Danforth*

Snow Entomological Museum, Department of Entomology, University of Kansas, Lawrence, KS 66045, USA

Received September 3, 1990 / Accepted June 4, 1991

Summary. In Perdita portalis, a ground nesting, commu- nal bee, males are clearly dimorphic. The two male morphs are easily distinguished based on head size and shape into (1) a flight-capable, small-headed (SH) morph that resembles the males of other closely related species and (2) a flightless, large-headed (LH) morph that pos- sesses numerous derived traits, such as reduced com- pound eyes, enlarged facial foveae and fully atrophied indirect flight muscles. The SH morph occurs exclusively on flowers while the LH morph is found only in nests with females. While on flowers, SH males are aggressive, fighting with conspecific males and heterospecific male and female bees, and they mate frequently with foraging females. Using artificial observation nests placed in the field, I observed the behavior of females and LH males within their subterranean nests. LH males are aggressive fighters; males attacked each other with mandibles agape, and male-male fights always ended in the death of one male. LH males are highly attentive to the repro- ductive behavior of females; they spend increasing amounts of time near open cells during cell provisioning, and mating only takes place immediately prior to ovipo- sition when females are forming the accumulated pollen and nectar into a ball. Based on larvae reared to adult- hood in the laboratory, the two male morphs occur in equal proportions. The behavior of males in closely re- lated species, especially P. texana, and the origin and maintenance of male dimorphism are discussed.

Introduction

Perdita (Macroteropsis) portalis is an unusual bee be- cause of the existence of two discrete male morphs that differ both in structure and behavior. One male morph is capable of flight, relatively small-headed and regularly

* Current address: Department of Entomology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560, USA

found on flowers visited by foraging females (SH morph). First discovered by Rozen (1970), the other, derived, morph is strikingly modified, as judged by com- parison with related bees, possessing an expanded head capsule, enlarged facial foveae, reduced compound eyes, elongate pointed mandibles, a three-pronged clypeus, and reduced thoracic musculature, which renders it flightless and restricted to a life within the natal nest (LH morph). Unlike most of the close relatives of Perdita portalis, which show continuous variation in head size, albeit with strong positive allometry, the males of P. portalis are truly dimorphic.

Flightless, macrocephalic males are known from a number of Hymenoptera: fig wasps in the genera Idarnes and Philotrypesis (Hamilton 1979; Murray 1987), ants in the genus Cardiocondyla (Stuart et al. 1987a, b; Kino- mura and Yamauchi 1987), and in a communal halictine bee, Lasioglossum (Chilalictus) erythrururn (Kukuk and Schwarz 1988). In all of these examples, flightless males engage in lethal male-male fighting in an enclosed space like a fig receptacle or a natal nest. Males that survive inter-male combat have access to a large number of re- ceptive females, with whom they mate. Based on the results presented in this study, male dimorphism in P. portalis has arisen in a similar context. Large-headed, flightless males remaining in the natal nest are capable of lethal male combat and mate repeatedly with the fe- male residents immediately prior to oviposition. Sexual selection, due to the advantages of last male mating, appears to be responsible for the evolution of this bi- zarre, macrocephalic male morph.

The nesting and foraging behavior of female P. porta- lis has been described elsewhere (Danforth, in prepara- tion), but can be summarized as follows. This species occurs from north-central Arizona east to the southwest- ern corner of New Mexico (Hidalgo, Grant and Luna counties) and south to Zacatecas state, Mexico (Timber- lake 1954, 1968). Females nest communally, with up to 29 adult females and 26 adult macrocephalic males per nest. Nests consist of subterranean tunnels leading to brood cells approximately 10 cm below the surface.

236

FEMALE BEHAVIOR:

New cell cons~lJctioNsleep

P R E - P R O V I S I O N

Feeding trip (35 mkO

~ra-nest mating (1-3 tines) __ Cell closure (13 rain3

" ~ • Egglay~g(18sec.)

__ Pollen ball formation (20 mix)

Cell pro "vBior~ (4-8 trips/cell, 27 rnin./~p)

JL~ P R O V I S I O N _ _ r , ~ POST-OVIPOSITION

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

TIME - -'- Fig. 1. Summary of typical female behavior throughout the day. Time throughout the day divided into three time periods: pre-provisioning, provisioning, and post-oviposition

Each female res ident cons t ruc t s one cell pe r d a y and fully p rov i s ions it wi th po l l en and nec ta r col lected f rom species o f Sphaeralcea (Malvaceae) . F o r a g i n g occurs be- tween 0900 and 1500 hours and each cell is fully p ro - v is ioned with four to e ight fo rag ing trips. Fo l lowing the last fo rag ing tr ip, females a d d nec ta r and shape the accu- m u l a t e d po l len into a spher ica l po l l en -nec ta r ball . A n egg is la id on the po l len bal l and the cell is c losed wi th soil. This cycle is r epea ted da i ly t h r o u g h o u t the pe r iod o f Sphaeralcea b loom, usua l ly f rom la te July t h r o u g h m i d - S e p t e m b e r (Fig. 1). Based on obse rva t ions o f fe- males wi th in ar t i f ic ia l nests, there is no evidence o f coop- e ra t ion or r ep roduc t ive divis ion o f labor .

Nests are large, wi th over 200 comple t ed cells pe r nest, and are re -used by several genera t ions . Therefore , active nests con ta in one to several matr i l ines , depend ing on the n u m b e r o f females in i t ia l ly found ing the nest. There is some over lap o f genera t ions because eggs la id ear ly in the nest ing season m a y p r o d u c e adul t s tha t emerge in the la t te r pa r t o f the season. However , the m a j o r i t y o f of fspr ing enter d iapause and r ema in as last- ins tar la rvae (p repupae ) t h rough the winter .

Materials and methods

This study was conducted at a locality 2 km north of Rodeo, Hidal- go County, New Mexico, where approximately 10 nests in 1987, 24 nests in 1988, and 33 nests in 1989 were located. Nests were found at other localities in both 1988 and 1989:30 km north of Hachita, Grant County, NM (1988) and 9.2 km east of Animas, Hidalgo County, NM (1989). All localities were mixed Chihuahuan desert/grassland habitat with scattered Sphaeralcea plants, the sole source of pollen and nectar for this species. Sphaeralcea subhastata predominates at the Rodeo locality and S. angustifolia at the Ani- mas locality.

Body weights were measured either "wet" (live or freshly killed) or "dry" (dried for at least 2 days at 55 ° C7). In order to calculate the relationship between wet and dry body weight, freshly killed bees were weighed, then dried and reweighed. Weights were mea-

sured to the nearest 0.01 rag. Linear measurements were made on a sample of 99 males, half of which were collected as adults in nests or on flowers, and the other half reared to adulthood from larvae or pupae collected in nests. Measurements (Fig. 2) included intermandibular distance (IMD), intertegular distance (ITD), fore- wing length (FWL), and hindtibia length (HTL). Measurements were made using a binocular microscope equipped with an ocular micrometer. Scaling relationships were calculated using major axis (Model II) regression of log-transformed data (Sokal and Rohlf 1981 ; LaBarbera 1989).

Nests were excavated by blowing a fine mist of talcum powder down the tunnels and carefully digging away the soil with a pen- knife. Care was taken to excavate nests early or late in the day or on cloudy days, in order to insure that all the resident bees were present. Larvae and pupae were placed in 96-well tissue cul- ture dishes for rearing in a laboratory incubator at 30 ° C.

In order to observe the behavior of males and females within nests under semi-natural conditions, observation nests, similar to those developed for the study of a social halictine bee, Lasioglossum (Dialictus) zephyrum, were used (Michener and Brothers 1971). The construction of observation nests is described in detail elsewhere (Danforth, in preparation). In brief, they consisted of a thin layer of soil sandwiched between two plates, 40.6 x 20.3 cm each, of transparent red pIexiglass. Bees were placed into these nests as pupae and, following emergence, readily accepted their artificial homes.

Unlike most studies of halictine bees in observation nests (Batra 1964, 1968 ; Stockhammer 1966; Kukuk and Schwarz 1987, 1988), nests were placed in the field so that bees could forage under natu- ral conditions. In order to do this, a hole, 1.5 m deep, 1.5 m wide, and 2 m long, was dug at the Rodeo locality and was covered with a 3/4-inch plywood top fitted with a trap door. Four slots, 1 x 20.3 cm each, were cut along one edge of the plywood top into which the uppermost edges of the observation nests could be fitted. The nests were supported from beneath by a wooden shelf.

In 1988 a total of seven nests were constructed and three were eventually accepted by at least one female and up to four LH males. In 1989 six nests were constructed and three were accepted by at least one female. LH males were deliberately excluded from the 1989 nests in order to observe female behavior in their absence; a situation like that found in newly initiated nests, prior to the emergence of the first LH males. The maximum number of adult females per nest was six and the maximum number of adult LH

A B

C ,"

D

Fig. 2A-D. Morphological measurements. A Anterior view male head, B dorsal view male mesosoma, C male hindtibia, D male forewing

males was four. Fifty-six cells were completed in the six successful observation nests. Nineteen of the 56 cells were observed for the complete provisioning cycle, and 4 were observed for part of the provisioning cycle.

During the periods of observation, usually from 0900 to 1500 hours, the behavior of all females and males in the observation

237

nests was recorded. For females, I recorded the times of departures from the nest, returns with and without pollen, cell construction, pollen ball formation, oviposition, and mating. For males, I contin- uously recorded their behavior and location within the nest (either at the nest entrance, in the main tunnel, in lateral tunnels leading to cells, or in cells) and when mating and male-male fighting oc- curred. In addition, I recorded the position and activity of each bee in the nest at 10-min intervals throughout the observation period (=scan sampling of Altmann 1974). Data were recorded with a handheld tape recorder and time intervals measured to the nearest second. A total of 69 h of observation were recorded (1988: 45.5 h; 1989:23.5 h).

Results

Male morphology and body size

Male P. portalis fall into two non-overlapping groups, most easily distinguished by head size. All the males collected as adults or reared from pupae or larvae could be unambiguously identified as either the LH or the SH morph based on intermandibular distance ( IMD; Fig. 3A). SH males had IMDs between 0.90 and 1.26 mm, while IMD ranged from 1.50 to 1.68 mm for the LH morph. This bimodal distribution of head widths does not result from the disappearance or death of males of intermediate head size, because males reared from larvae or pupae (Fig. 3 B) show the same degree of di- morphism as males collected as adults (Fig. 3 C).

There is some suggestion that the sample of SH adult males collected on flowers ( 2 = 1 . 1 6 + 0 . 0 1 ; n=30) may have slightly larger heads than the sample of SH males reared from larvae and pupae (2=1.11_+0.02; n--22), although the difference falls just short of significance (Fig. 3B, C; t=2 .00 ; P=0.052). The two samples of

1 5 A 1 0 B - 1 0 - C"

! 1 2 - - 8 8 ,-- r -

~- _ i |

z >- ~- uJ o "~ 9 - 7 z 6 z 6 - - O' / w w

0c 0 - 0 -- U. W W

3 2 - 7 2

i v . 7 . T , , , , , . . 7

IMD (mm) IMD (mm) IMD (mrn) Fig. 3A-C. Histograms of intermandibular distance ( IMD) in mm for the males of P. porta/ix. A shows data for all males (n=99), B shows data for lab-reared males (n=45), and C shows data for males collected as adults either on flowers or in nests (n=54). Means indicated along the abscissae with closed circles

238

A B C

>- 0 z uJ

0 uJ n- It.

1 5

1 2

1 0

)- 0 z 6 LU

0 UJ n- u. 4

r i (@v i

DRY WEIGHT (mg) DRY WEIGHT (mg)

10"

>.. 0 z 6 LU

0 LU n-" " 4

¢; .,9

DRY WEIGHT (mg)

Fig. 4A-C. Histograms of adult dry body weight. A small-headed males (n= 52), B macrocephalic males (n=46), and C females (n= 36). Means indicated along the abscissae with closed circles

Table 1. Morphological comparison of small and large-headed males and females. Mean _+ SE (N), and range. See Fig. 2 for mea- surements

Small- Large- headed males headed males P

Mb (rag) 0.87 1.26 < 0.001 _+ 0.03 (52) _+ 0.04 (46) 0.45-1.27 0.83-1.79

IMD (ram) 1.14 1.57 <0.001 _+0.01 (52) +0.01 (47) 0.88-1.28 1.47 1.68

ITD (mm) 0.92 0.88 0.0037 +0.01 (50) _+0.01 (47) 0.74-1.06 0.80-0.96

FWL (ram) 2.82 2.48 <0.001 +0.02 (49) _+0.01 (39) 2.48-3.04 2.36-2.60

HTL (ram) 0.88 0.86 0.014 ±0.01 (52) ___0.004 (46) 0.74-0.98 0.80-0.93

Females P

Mb (rag) 1.36 <0.001 (vs small- _+ 0.04 (36) headed males) 1.00-1.80 0.04 (vs large-

headed males)

LH males, those collected as adults in nests (2= 1.57_+ 0.01 ; n = 24) and those reared from larvae or pupae (2 = 1.58_+0.01; n=23), differ very little in head width (t= 0.83; P = 0.41).

LH and SH males differed significantly in all other variables measured: body weight (Mb; Fig. 4A, B), inter-

Table 2. Coefficients of variation for males and females. P values give the probability that parametric variances are equal

Small- Large- P headed males headed males

M b 20.58% IMD 7.97% ITD 7.32% FWL 4.46% HTL 5.90%

Females

M b 15.43 %

20.10% <0.001 3.15% <0.001 4.29% <0.001 2.38% <0.001 3.45% < 0.001

(P<0.25, n.s. for both comparisons)

tegular distance, forewing length (Fig. 8), and hind tibia length (Table 1). LH males are significantly larger in body weight and IMD, but significantly smaller in ITD, FWL, and HTL.

The SH males are also significantly more variable than the LH males (Table 2). For all linear measurements and dry body weight, the variances were significantly greater for the SH males than for the LH males. The variance in female dry body weight did not differ signifi- cantly from the variance of either SH or LH males.

Figure 5 shows the relationships between the log- transformed linear measurements and the log of dry body weight for the two male morphs. The allometric equations, calculated separately for each morph, are shown in Table 3. Since these rleationships are all be- tween a length and a mass, the null hypothesis of isometry, shown in Fig. 5 with the dashed line, is a slope of 0.33. For 1MD and ITD, the SH morph does not differ from isometry (Table 3); however, for FWL and

239

0 . 2 0

0 . 1 0

0.00

0.00.

a

- 0 . 0 6 "

- 0 . 1 2 "

o OO o ~ o o ~ o o ~

A o ~'oo o C o o o o ~ 8 3 ~

" . o 0 , .t~.-("o--C~/~ ~ 0 . 4 5

U-

: ". o 4 o

. . J . J

• 0.35

I I I I I I

- 0 . 3 6 - 0 . 2 4 - 0 . 1 2 0 .00 0 . 1 2 0 . 2 4

log Mb ( r a g )

0 . 5 0

B / ~ / "

• ..C %, .;-j ~ - 0 .04 -

. ' o 8 / • 00 • o 0o o o • o / ~ o

"•e ~ e~ o C o o ~ o 5 ~ So,-"~-o" • • ooo-~o'oooe~ o o o ~ -0.08" -" o~~o-; o o

/ ~- - 0 . 1 2 " r e

I I I I I I

- 0 . 3 6 - 0 . 2 4 - 0 . 1 2 0 .00 0 .12 0 . 2 4

log Mb (mg)

C / / ~ / ~

% 0 ~ o~ e 'o. . . .o o-"g-

• ~ o/. .o o• o • ~ •

/ o 0 o ~ o o

i o o / o /i / i o

/ //

I I I I ~ I

-0.36 -0.24 -0.12 0.00 0.12 0.24

log Mb (rag) / /

/

/ /

• / / "

D " ".~:, ,'.,~-" • ,, o /

• o o o ~ • • to /o o • • • o • " ; ~ " • e ~ o oo o• %0 o~o ° /

/ ,," o• ~_qo~o6~oo" o ~ o • - / 0 , , ~ ,o d o oo] o o

, ~ o o 2 , • o o

• I Co /

" • i / I /

/ /

, /

I I I I I f

- 0 . 3 6 - 0 . 2 4 - 0 . 1 2 0 . 0 0 0 . 1 2 0 . 2 4

log Mb (rag)

Fig. 5A-D. Bivariate plots of linear measurements (log-transformed) versus the log of dry body weight. Closed circles indicate SH males and open circles indicate LH males. Dashed lines indicate the null hypothesis of isometry (a slope of 0.33) and the solid lines indicate the major axis regression lines for each group

Table 3. All•metric equations, correspond- ing to regression lines shown in Fig. 5, re- lating linear measurements to dry body weight for both male morphs. P indicates probability that the observed all•metric coefficients, the exponents in these equa- tions, are indistinguishable from the null hypothesis of isometry, 0.33, using a X 2 test with 1 degree of freedom (Jolicoeur 1963b; Anderson 1963)

Small-headed male P Large-headed male P

IMD = 1.20 Mb 0"32 < 0.9 IMD = ] .55 Mb 0'09 < < 0.001

ITD=0.95 Mb °'27 <0.5 ITD=0.85 Mb °'12 < <0.001

FWL=2.88 Mb °'15 <0.00] FWL=2.45 M b -0.004 < <0.001

HTL=0.91 Mb °'19 <0.001 HTL=0.85 Mb T M < <0.001

HTL, the slope of the major axis regression line is signifi- cantly less than 0.33. For the LH morph, all four linear measurements scale negatively all•metrically. In fact, the regression lines, with slopes ranging f rom - 0 . 0 0 4 to 0.12, are nearly horizontal, indicating that for LH males a change in body mass results in very little change in linear measurements. This is consistent with the low vari- ability in these linear measurements among L H males (Table 2). Figure 5 illustates a morphological t radeoff between investment in head size and investment in non- head body size: L H males, al though they are significant- ly heavier overall, have disproportionately small bodies for their size, as measured by the three thoracic measure- ments. The SH morph, on the other hand, has a dispro- port ionately large thorax, relative to total weight.

The differences between L H and SH males in external morphology are associated with differences in internal morphology, primarily in flight and mandibular muscu- lature. In both male morphs, the bulk of the head cap-

sule is taken up by the mandibular adductor muscles, which lie behind the brain and paired mandibular glands. The adductor muscles originate over the inner surface of the head capsule to the level of the median ocellus in front, down the sides of the head capsule, and to the level of the occipital foramen posteriorly (Fig. 6 A, D). Portions of the adductor muscles also orig- inate on a cephalic lamella lying along the midline of the head f rom the anterior ocellus to the occipital fora- men, and also on the lateral surfaces of the tentorium. Because the region of the head dorsal to the •cell• is expanded in the L H males, the surface area available for adductor muscle a t tachment is greatly enlarged.

The mandibular abductor muscles originate on the posterior and lateral areas of the head capsule and do not differ conspicuously in size between the two morphs (Fig. 6 B, E).

The two morphs differ in the external structure of the head capsule, especially on the posterior surface. The

240

A m. add. B *^- t . C / occ. for.

/ hyp. car.

abd / ~ ( ~ / ~ x ' ~ / /prob. foss" o

/ -x ; / gen. proj.

D E

Fig. 6A-F. Head morphology of the two morphs in P. portalis. A-C LH morph; D--F SH morph. A, B, D, E show anterior view of male head; C, F show posterior views, m.add.=mandibular adductor muscle, m.abd. = mandibular abductor, tent. = tentorium,

F

occ.for.=occipital foramen, hyp.car.=hypostomal carina, prob.- foss. =proboscidial fossa, gen.proj. = genal projection. Scale bar = 1.0 mm

c

dvfm.

E

. [ofml / dvfm. fa. bo.

Fig. 7A-J. Thoracic morphology of the two male morphs. A-D lateral views of SH male thorax, E-F lateral views of LH male thorax, G, H SH male second phragma, I, J LH male second

I j

phragma, lofm. =longitudinal indirect flight muscle, dvfm. =dor- so-ventral indirect flight muscle, 2Ph. =second phragma, vnc. = ventral nerve chord, fa.bo. =fat body. Scale bar= 1.0 mm

SH m o r p h has a p r o b o s c i d i a l fossa s imi lar to males o f o the r species o f Perdita (Fig. 6 F ) ; however , in the L H m o r p h the h y p o s t o m a l ca r inae are p r o d u c e d in to thin lamel lae tha t a rch over the p robosc id i a l fossa (Fig. 6C) ,

fo rming a r o o f tha t pa r t ly encloses the w i t h d r a w n labio- max i l l a ry complex .

The two m o r p h s also differ s t r ik ingly in thorac ic mor - pho logy . In SH males the do r so -ven t r a l and long i tud ina l

Macrocephalic male

Fig. 8. Wing morphology of the two male morphs. Scale bar= 1.0 mm

flight muscles are large and conspicuous (Fig. 7 C, B, D). However, in the LH morph these muscles are entirely lacking and fatty tissue is found in their place (Fig. 7 F). The curvature of the scutum, the site of attachment of the flight muscles (Fig. 7A, E), is reduced, and the meso- thoracic phragma, another site of attachment for the longitudinal flight muscles, is reduced in the flightless morph (Fig. 7G, H vs I, J). For a description of the flight mechanism in bees, see Snodgrass (1956). Al- though the flight musculature is lacking in the LH morph the wings are not conspicuously modified, al- though they are somewhat smaller (Fig. 8; Table 1).

The relationship between dry and wet body weight was determined for 13 males (11 SH and 2 LH) and 20 females. For males the regression equation is Mb (wet)=--0.027+2.85 Mb (dry) (P<0.00), and for fe- males is Mb (wet)=--0.161+2.76 Mb (dry) (P<0.00). The two LH males do not deviate significantly from the regression line for all males combined. The propor- tion of body weight made up of water is significantly higher for males (2=64.3_+0.7%; n=13) than for fe- males (2=61.9_+0.7%; n=20; P=0.036 by Mann- Whitney U-test), but LH and SH males did not differ.

In summary, the two male morphs are distinct mor- phologically in the mean values for both weight and linear measurements, the degree of morphological varia-

241

tion, the scaling relationships between several linear measurements and body weight, and cuticular and mus- cular morphology. As in dimorphic thrips (Crespi 1988), beetles (Eberhard 1982) and fig wasps (Hamilton 1979), the dimorphism in P. portalis reflects a tradeoff between investment in morphological features associated with flight (thoracic musculature, wing length) and morpho- logical features associated with male-male fighting (man- dibular musculature, head armament).

Behavior of SH males on flowers

The dimorphism described above corresponds to behav- ioral diphenism as well. LH males are found exclusively in nests and are flightless. The SH morph is found almost exclusively on flowers, but occasionally, recently eclosed SH males can be found in nests. Detailed observations of SH males on flowers for up to 30 min indicated that they pursue a relatively stationary, sit-and-wait, mating strategy. Approximately half of the 32 males observed remained on the same flower for more than 30 min. While on flowers, males are attentive to the presence of other insects. When a conspecific female landed on a flower, the resident male either pounced on her imme- diately and mating ensued (59/88 male:female en- counters; 67%), or the female left the flower (29/88 male: female encounters; 33 %). While on flowers, males mated frequently - up to seven matings were observed per half hour. Mating lasted from 10 s to over 2 min (2= 1.18-t-0.05 min; n= 59).

When insects other than conspecific females landed on a flower occupied by a male P. portalis, males be- haved aggressively. They commonly attacked both male and female P. latior, a closely related but larger Sphaer- alcea visitor, various bombyliid flies, and, especially ag- gressively, other P. portalis males. On six occasions, male P. portalis landed on flowers occupied by a conspecific male and in every instance fighting ensued. The fights were violent, with the two males rapidly tumbling around the cup-shaped corolla, locked together with their mandibles. Fights lasted up to 40 s and ended when one male flew off. Based on censuses of open Sphaeral- cea flowers at the Rodeo locality during the time of female foraging, from 3 to 12 SH males and up to 12 females were present per 50 open flowers.

Macrocephalic male nest occupancy

Perdita portalis is a communal bee with from 2 to 29 adult females and up to 26 adult males per nest (n = 9 active nests excavated). Two nests contained SH males, probably recently emerged bees that had not yet left their natal nest. No SH males remained in observation nests.

There is evidence that the number of LH males per nest is correlated with the number of females per nest. Figure 9 shows the relationship between the number of males and females in active nests excavated in 1987- 1989. Nests 89-17 and 89-50 appeared, based on the

242

3 0

20. CO LU ..I <

d 10 ~ ~ e e s Z 8 -5

B8-12 • " 0 _ 8 7 - I

0 / 89-11 0 " 8 9 _ _ 5 1 & 8 9 - 5 0 8 9 - 1 7 •

I I I I I I

5 10 15 20 25

NO. F E M A L E S / N E S T

Fig. 9. Relationship between number of macrocephalic males and number of females per nest (n = 9 nests fully excavated). Solid line indicates equal numbers of males and females. Old, persistent nests are shown with closed circles; newly initiated nests are shown with triangles

uJ

p.- O I- I- Z LU n- LU EL

8 0 '

7 0 '

6 0 .

5 0 '

4 0 '

3 0 '

20 "

10 .

MALE INTRA-NEST BEHAVIOR

I::::

I:i:i

] Entrance [ ] Lateral

[ ] Maintunnel • Cell

IIi n 1

TIME I I i i I i PERIOD P R E - P R O P R O V . / j POST-OVIP.

Cell provisioning begins Egg laid

Fig. 10. Data on male behavior within nests based on scan sampling (Altmann 1974). Bars indicate the percentage of time spent in var- ious locations in the nest (e.g., "Entrance" means at the entrance to the nest) Overall time is broken up into three time periods, explained in text (cf. Fig. 1). Data for single- and multiple-male nests combined

small number of prepupae, to be newly initiated. In the remaining seven mature nests the number of males and females per nest appear roughly equal (rho = 0.977; P < 0.01, Spearman's rank correlation coefficient).

Behavior of macrocephalic males within observation nests

Figures 10 and 11 summarize the data on LH male be- havior based on 50 male hours of observation of five

s O

1 0 0 -

90 -

80 -

70 -

60 _

50 -

40 -

30 -

20 -

10 -

90 -

80 --

70 -

60 -

4 O

3 O

2 0 - -

1 0 - -

r-3

Single male nests

] Entrance

] Tunnel

Cell

] Mating

[ ] Fighting

i - - i m

Multiple male nests

i!ii

iiii

!:!::

TIME PERIOD [ I ~ I ~ J P R E - P R O V j / ~ / / ~ PROV. POST-OVIP,

Cell provisioning begins Egg ~id

Fig. 11. Data on male behavior within nests based on continuous record. Bars indicate percentage of time spent in various activities. " E n t r a n c e " : bee was out of sight of the observer and therefore within 3-4 cm of the nest entrance; "Tunnel": bee was resting or moving about in the main and lateral tunnels; "Cell": bee was in an open cell; "Mating": the bee was observed mating in the main or the lateral; and" Fighting" : males were observed fight- ing in the main or the lateral. The data for single- and multiple- male nests are presented separately. Overall time is broken up into three time periods, explained in text (cf. Fig. 1)

males in three observation nests. Observation periods were broken up into three catagories: (1) pre-provision- ing, when open, unprovisioned cells were present in the nest; (2) provisioning, when open, partially provisioned cells were present and females were actively provisioning cells; and (3) post-oviposition, when all cells had been closed or abandoned and females were no longer actively provisioning (Fig. 1). Figure 10 shows the data based on scan sampling and Figure 11 presents the data based on continuous monitoring.

Males spent very little time at the nest entrance over- all: 5.5% based on scan sampling and 2.7-2.8% based on continuous monitoring. Each visit to the nest en- trance lasted 47.4_+9.3 s (n=96) , on average. There is little fluctuation in the percentage of time spent at the nest entrance during the different time periods.

Although LH males spent the majority of time in the main and lateral tunnels, there are significant changes over the course of a day in the percent of time spent in various locations within the nest. For example, in the transition from pre-provisJoning to provisioning

243

time periods, for the scan sampling data (Fig. 10), time spent in lateral tunnels leading to open cells (18.4% to 39.9%) and cells (0.8% to 19.1%) increased with a corre- sponding decrease in time spent in the main tunnel (76.0% to 34.0%; )~2--96.93, df=3, P<0.001). Follow- ing oviposition, males returned to pre-provisioning types of behavior, again spending the majority of their time in the main tunnel (66.7%) and less time (33.3%) in lateral tunnels (Z 2 = 11.07, df= 3, P<0.025).

Similar changes in the behavior of LH males are evi- dent in the data based on continuous records, at least for single-male nests: time spent in the main and lateral tunnels drops from 98.0% to 75.1%, and time spent in open cells increases from 0.0% to 18.7% (Fig. 11). This trend is not evident for multiple-male nests, probab- ly because of increased male-male fighting in multiple- male nests, which took place mostly in the main tunnel and laterals.

During the period of cell provisioning, males behaved as though they were guarding entrances to laterals lead- ing to partially provisioned cells. While in laterals, they were always facing away from the open cell, and were usually located at the junction of the lateral and main tunnels. When encountering bees that were descending the main tunnel, LH males often acted aggressively to- ward them, at least initially. If the descending bee was a female, she was allowed to pass, but when other LH males were encountered, they were attacked.

Males spent on average 3.1 _+0.3 s (n= 59) in an open cell during each visit. When entering partially pro- visioned cells containing loose clumps of pollen, males could frequently be seen lifting up the pollen loads in their mandibles. LH males were never seen receiving food from female nest-mates through oral trophallaxis, as in Lasioglossum erythrurum (Kukuk and Schwarz 1988), so it is likely they are feeding on the pollen, or the small amounts of nectar in the pollen loads, during these brief visits to partially provisioned cells.

Intra-nest mating only took place during pollen ball formation, just prior to oviposition (Fig. 11). After fe- males had completed their last foraging trip of the day, they began to form the accumulated pollen into a pollen ball (Fig. 1). Pollen ball formation took 19.93 _+ 0.93 min (n = 16), and approximately 9 min after the initiation of pollen ball formation, females invariably left their cells and entered the lateral tunnel or, occasionally, the main tunnel. If a LH male was present, mating always took place during this departure from the cell (9 of 9 times). In nests that lacked LH males, females simply groomed during this period and then returned to the cell (10 of 10 times) and continued pollen ball formation.

In nests with LH males, from one to two additional copulations occurred before the egg was laid, but unlike the first mating, the subsequent matings were initiated when the male entered the cell and pulled the female out into the lateral tunnel with his mandibles. These later, male-initiated, matings were recorded in eight of the nine cells I observed completely provisioned. No mating was observed within nests at any other time ex- cept during pollen ball formation, and for every cell observed through the completion of provisioning, fe-

males mated at least once with a resident LH male, when present, prior to egg laying.

Based on the continuous records of male behavior, males spent 1.2% of their time mating in single-male nests and 0.4% of their time mating in multiple-male nests. Intra-nest matings lasted on average 1.96___ 0.13 min (n = 16).

Male-male fighting within observation nests

Males fight with each other when they are present in the same nest (Fig. 11). Overall, males spent 3.4% of their time fighting; however, this value gives no indica- tion of the intensity of these battles. LH males behaved aggressively when they encountered another male in the nest. During fights, males raised their heads and held their mandibles agape, and the two opponents grasped each other by the head and pushed, pulled, and twisted for, on average, 45.6 _+ 6.9 s (n = 51). One male was killed after 13 encounters over the course of 3 h. On the follow- ing day, another male was killed after 46 battles in 5 h. Over the next 2 days two more LH males eclosed and both were later found dead. Males apparently died as a result of wounds to the neck region or the membranous roof of the proboscidial fossa (Fig. 6 C, F). The function of the elongate genal projections and the well-developed hypostomal carinae in LH males appears to be to protect this vulnerable region from the opponent's mandibles during intra-nest battles.

Although these observations suggest that LH males cannot coexist in the same nest, as many as 26 LH males were found in natural nests. The correspondence be- tween the number of females and the number of LH males (Fig. 9) suggests that males may be able to tolerate other males in the same nest as long as the nest is large enough to allow each male access to at least one female. Perhaps in larger nests the frequency of male-male en- counters is reduced, and thus a number of males can occupy the same nest. Dead LH males were found while excavating two natural nests, and one dead macroce- phalic male was found next to a nest entrance, indicating that males also fight to the death under natural condi- tions.

Behavior of macrocephalic males in natural nests

In 1988, data were collected on the rate and duration of LH male visits to nest entrances, in natural nests. LH males were easily identified because they differ con- spicuously from SH males and females in the color and punctation of the head. LH males spent from 1 to 314 s at the nest entrance on each visit (2= 13.59_+4.26 s; n= 79 visits). A LH male appeared at the nest entrance at a rate of 4.26 visits/h. Furthermore, of the 23 natural nests studied, but not excavated, LH males were seen in all but 5. This value may underestimate the occurrence of LH males because they may have been present, but not seen, in these 5 nests. Considering both excavated (n = 4) and observed (n = 23) nests, LH males were pres- ent in at least 81% (21/27) of the nests.

244

Sex and morph ratio

The sex ratio of 167 adults reared from nests excavated in 1988 and 1989 was 55% male, 45% female (not signif- icantly different from a 50:50 sex ratio; Z2= 1.73, P < 0.5). Of the 80 males reared and identified to morph, 52.5% were LH males and 47.5% were SH males (not significantly different from a 50:50 morph ratio; 2~2= 0.20, P<0.9).

Discussion

Origin of male dimorphism in P. portalis

P. portalis belongs to a monophyletic group of 30 species that includes the subgenera Macrotera, Macroteropsis, Macroterella, and CockereIlula (Timberlake 1954; Dan- forth, in preparation). Based on the phylogeny of this group and information on the nesting biology and male mating behavior of other species in the group [P. (Mac- roteropsis) latior, Rozen, personal communication; P. (Macrotera) texana, Neff and Danforth, in preparation; P. (Macrotera) sp. nov. Rozen, personal communication; P. (Cockerellula) opuntiae, Custer 1928, Bennett and Breed 1985; and P. (Macroterella) mellea, Rozen, per- sonal communication], the LH male in P. portalis is almost certainly a derived trait for this species. All other species studied to date, except P. mellea, show continu- ous variation in male head size. However, in these species male head size shows positive allometry such that larger males have disproportionately large heads (e.g., P. tex- ana; Neff and Danforth, in preparation). Furthermore, in P. texana, and possibly others, although all males are flight-capable, some males periodically enter nests for extended periods of time, during which mating may take place. Male dimorphism appears to have arisen in- dependently in P. mellea and P. portalis from a condition of continuous variation in male head size and shape, like that seen in male P. texana. What factors explain the origin of male dimorphism in P. portalis?

Kukuk and Schwarz (1988) list three hypotheses for the existence of male dimorphism in Lasioglossum eryth- rurum: (1) LH males may result when females mistakenly lay a male egg on a large pollen ball originally destined for a female offspring. This hypothesis assumes that the LH males are developmental anomalies that have neither their origin nor their maintenance explained by selection. The other two hypotheses invoke selection to explain the origin and persistence of the LH morph: (2) LH males are a "soldier caste" resulting from kin-selection favoring male nest guarding (Houston 1970; see Bartz 1982 for a theoretical discussion of the evolution of male workers) and (3) LH males result from intra-sexual selec- tion (Huxley 1938) favoring increased fighting ability, which enables males to monopolize females in a commu- nal nest by killing or excluding conspecific male competi- tors.

The first hypothesis is plausible for P. portalis because females are significantly heavier than SH males (Fig. 4), but very similar in body weight to the LH morph. In

addition, strong positive allometry in male head size is present in the species of Perdita closely related to P. portalis (Danforth, in preparation). A significant corre- lation between provision mass and adult body weight among individuals of the same sex has also been ob- served in several hymenopteran species (Dow 1942; Klostermeyer et al. 1973; Plowright and Jay 1977; Cross et al. 1978; Cowan 1981 ; Jayasingh and Taffe 1982; Pla- teaux-Qu6nu 1983; Brockmann and Grafen 1989).

However, a number of facts cast doubt on the ade- quacy of oviposition °'mistakes'' in explaining the ex- istence of the LH morph. First, LH males occur at a very high frequency; more than 50% of the males pro- duced are the LH morph. Second, the dimorphism is absolute, with no males of intermediate morphology. Finally, the behavior of the LH males within nests, spe- cifically lethal male combat and mating immediately prior to oviposition, suggests that they are unlikely to be developmental anomalies.

Hypothesis two, that LH males act as guards protect- ing the nest from intruders and parasites, is not well supported. Nest parasites or predators were never dis- covered in nest excavations, and only two cases of nest parasitism were observed in artificial nests (Danforth, in preparation). In addition, both in observation and natural nests, LH males spent only a small percentage of their time at the nest entrance, which is where female guards in halictine bees normally reside (Batra 1964; Brothers and Michener 1974; Abrams and Eickwort 1981).

The third hypothesis, that the LH morph has arisen as a result of male-male competition for females in the nest, is the strongest hypothesis and has been proposed for the origin of flightless males in other Hymenoptera (Hamilton 1979). Both the intensity of male-male com- bat in observation nests and the occurrence of dead LH males in natural nests indicate that inter-male combat is an important selective factor for males that remain in the natal nest. This hypothesis is further supported by: (1) the increased time spent by LH males in the vicinity of open cells during cell provisioning (Figs. 10, 11), (2) the behavior of LH males toward other bees in the nest during cell provisioning, and (3) the positive correlation between the number of LH males and fe- males in nests (Fig. 9). Finally, the repeated matings im- mediately prior to egg laying indicate that males are attentive to the reproductive behavior of females. If there is a last-male mating advantage in this species, the mating behavior of LH males would insure paternity of female-producing eggs laid by any females whom they inseminated prior to egg laying. From the perspective of LH males, mating with females prior to the beginning of pollen ball formation is wasted effort because females still have the opportunity to mate with SH males on flowers during subsequent foraging trips.

The role of sexu,al selection is dependent upon the degree of sperm precedence in P. portalis. Among bees, paternity studies have been conducted only for Apis mel- Iifera, for which there was no evidence of non-random sperm utilization by multiple-mated queens based on al- lozyme data (Page and Metcalf 1982). However, high

Table 4. Summary of female nest occupancy patterns and male variation in species closely related to P. portalis

P. sp. nov.

P. texana

P. latior

P. opuntiae

P. mellea

P. portalis

245

No. Male females/ variation nest

Reference

1-2 polymorphic J. Rozen, personal communication

1-28 polymorphic Neff & Danforth, in preparation

1-8 polymorphic J. Rozen, personal communication

1-38 polymorphic Bennett and Breed 1985

4-15 dimorphic J. Rozen, personal communication

2-29 dimorphic

levels of sperm precedence have been detected for other Hymenoptera: two species of Polistes (Metcalf 1980), Dahlbominus fuscipennis (Wilkes 1966) and Nasonia vitri- pennis (Holmes 1974; for reviews of sperm precedence in insects, see Eickwort and Ginsburg 1980; Parker 1970; and Starr 1984). Since multiple mating is common in Perdita portalis, as in other panurgine bees (e.g., No- madopsispuellae, Rutowski and Alcock 1980; Calliopsis andreniformis, Shinn 1967; and other species of Perdita, Danforth 1989), and sperm is stored in the spermatheca, some degree of sperm precedence seems likely.

If intra-sexual selection explains the evolution of the LH morph in P. portalis (and P. mellea), why has di- morphism not arisen in similar, closely-related species? One factor favoring the evolution of flightless males, which has been suggested for fig wasps (Hamilton 1979) and ants (Kinomura and Yamauchi 1987), is the co- occurrence of multiple, reproductive females. The number of co-habiting, receptive females represents the magnitude of the potential reproductive payoff to a male that can gain access to those females. Table 4 shows data on female nest occupancy for P. portalis and the five closely related species for which these data are avail- able. Although P. portaIis and P. meIlea are among the species with the highest number of females/nest, other species, which lack LH males, also have large nest popu- lations (e.g., P. opuntiae). Although P. texana may con- tain a maximum of 28 females/nest, such large nests are very rare and the majority of nests contained less than 4 females. These data do not strongly support this hypothesis. The hypothesis is further weakened by the fact that many males occur together in nests, in propor- tion to the number of adult females (Fig. 9).

A second factor that may favor LH males is the long- term persistence of nests. In P. texana the vast majority of females appear to found new nests each year (Neff and Danforth, in preparation). In contrast, natal nest re-use appears to be far more prevalent in P. portalis, judging by the size of nests and their persistence year to year (Danforth, in preparation). Frequent nest re-use by females would favor the evolution of flightless, LH males by eliminating the need for males to leave the natal nest in search of receptive females.

Finally, differences in the degree of non-random sperm utilization may underlie differences in male mat-

ing behavior. A high level of sperm precedence would impose strong selection on males that could gain access to females immediately prior to egg laying.

Maintenance o f dimorphism

Assuming this dimorphism is a stable one, and that the plesiomorphic SH morph is not being eliminated by se- lection, one may ask what factors maintain both male morphs in this species. The answer to this question will depend on whether the two morphs are equal or unequal in fitness (Austad 1984; Dominey 1984). The two morph fitnesses could be equal for several reasons. First, there may be no sperm precedence. Alternatively, fitness may be frequency-dependent. Neither of these hypotheses seems likely.

The dimorphism may persist in spite of unequal fit- ness if relative morph fitness varies in space and/or time (Dawkins 1980; Austad 1984). Temporal and spatial variation in the proportions of newly founded nests, which lack LH males, and old persistent nests, which contain LH males, could contribute to the maintenance of the dimorphism. Newly founded nests must lack LH males prior to the emergence of the oldest brood. Devel- opment from egg to adult takes at least 21 days in P. portalis (Danforth, in preparation). Therefore, for at least a 3-week period, females in newly initiated nests will be mating exclusively with SH males. Variation in the number of old versus new nests could arise at one locality over the course of a season (nests lacking LH males are most prevalent early in the season), or between localities due to the mix of new and old nests, or popula- tion density.

The two male morphs may also persist in the popula- tion because of trade-offs imposed on females by the difference in relative "cost" of the two morphs. The question of how females allocate their total male repro- ductive effort between LH and SH males is analogous to the question of how females make sex ratio allocation decisions. Fisher's model for sex ratio allocation predicts that selection acting on the offspring generation can alter decisions made by parents in their output of sons and daughters (Fisher 1958; Dawkins 1980), As in sex ratio allocation, in which the two sexes are not necessarily

246

equal in cost to the parent , the two male morphs in P. portalis are p robab ly unequa l in cost to the parent . Since LH males are significantly heavier, they are prob- ably significantly more costly in t ime and energy to pro- duce (Danfo r th 1990). A l though L H male fitness may exceed SH male fitness, the two morphs may persist together because, f rom the perspective of ma te rna l fit- ness, a few " c o s t l y " LH males may equal m a n y " inex- pens ive" SH males (Alcock et al. 1977). U n d e r this scen- ario, SH males mus t adop t a strategy of " m a k i n g the best of a bad j o b " (Dawkins 1980; Eberhard 1982).

Acknowledgements. I am grateful to C.D. Michener and J.G. Ro- zen, Jr. for their advice and help during this study. Without Dr. Rozen's generous help in the field this project would never have been started. I thank the resident directors of the American Muse- um of Natural History's Southwestern Research Station, Wade Sherbrooke, Pare Limberger, and Kristina Schwartz, and the vol- unteers, Emily Dickinson, James A. Davidson, David Domingo, and Ernie Luikart for their help in collecting data. This study was supported by two Theodore Roosevelt Memorial Grants from the American Museum of Natural History and a Pre-doctoral Fel- lowship from the Smithsonian Institution. G.C. Eickwort, J. A1- cock, J. Brockmann, P. Kukuk, C.D. Michener and J.G. Rozen, Jr. commented on earlier versions of this paper. This is Contribu- tion No. 3035 from the Department of Entomology and the Snow Entomological Museum, University of Kansas, Lawrence.

References

Abrams J, Eickwort GC (1981) Nest switching and guarding by the communal sweat bee Agapostemon vireseens (Hymenoptera, Halictidae). Insectes Soc 28:105-116

Alcock J, Jones CE, Buchmann SL (1977) Male mating strategies in the bee Centris pallida Fox (Anthophoridae: Hymenoptera). Am Nat 111:145-155

Altmann J (1974) Observational study of animal behavior: sam- pling methods. Behavior 49 : 227-265

Anderson TW (1963) Asymptotic theory for principal component analysis. Ann Math Star 34:12~148

Austad SN (1984) A classification of alternative reproductive be- haviors and methods for field-testing ESS models. Am Zool 23 : 309-319

Bartz SH (1982) On the evolution of male workers in the hymenop- tera. Behav Ecol Sociobiol 11 : 223-228

Batra SWT (1964) Behavior of the social bee, Lasioglossum zephyr- urn, within the nest (Hymenoptera: Halictidae). Insectes Soc 11:159 185

Batra SWT (1968) Behavior of some social and solitary halictine bees within their nests: a comparative study. J Kans Entomol Soc 41 : 120-133

Bennett B, Breed MD (i 985) The nesting biology, mating behavior, and foraging ecology of Perdita opuntiae (Hymenoptera: An- drenidae). J Kans Entomol Soc 58:185-194

Brockmann HJ, Grafen A (1989) Mate conflict and male behaviour in a solitary wasp, Trypoxylon (Trypargilum) politum (Hyme- noptera: Sphecidae). Anita Behav 37: 232-255

Brothers D J, Michener CD (1974) Interactions in colonies of primi- tively social bees III. Ethometry and division of labor in Lasiog- lossum zephyrum (Hymenoptera: Halictidae). J Comp Physiol 90 : 129-168

Cowan DP (1981) Parental investment in two solitary wasps An- cistrocerus adiabatus and Euodynerus foraminatus. Behav Soc Sociobiol 9:95 102

Crespi BJ (1988) Adaptation, compromise and constraint: the de- velopment, morphometrics, and behavioral basis of a fighter- flier polymorphism in male Hoplothrips karnyi (Insecta: Thy- sanoptera). Behav Ecol Sociobiol 23:93-104

Cross EA, Mostafa A E-S, Bauman TR, Lancaster IJ (1978) Some aspects of energy transfer between the organ pipe mud-dauber Trypoxylon politum and its araneid spider prey. Environ Ento- mol 7:647-652

Custer CP (1928) The bee that works stone; Perdita opuntiae Cock- erell. Psyche 35:67-84

Danforth BN (1989) Nesting behavior of four species of Perdita (Hymenoptera: Andrenidae). J Kans Entomol Soc 62:59-79

Danforth BN (1990) Provisioning behavior and the estimation of investment ratios in a solitary bee, Calliopsis (Hypomacrotera) persimilis (Cockerell) (Hymenoptera: Andrenidae). Behav Ecol Sociobiol 27:159-168

Dawkins R (1980) Good strategy or evolutionarily stable strategy? In: Barlow GW, Silverberg J (eds) Sociobiology: beyond na- ture/nurture? Westview, Boulder, pp 331-367

Dominey WJ (1984) Alternative mating tactics and evolutionarily stable strategies. Am Nat 24:385-396

Dow R (1942) The relationship of the prey of Sphecius speciosus to the size and sex of the adult wasp. Ann Entomol Soc Am 35:310-317

Eberhard WG (1982) Beetle horn dimorphism: making the best of a bad lot. Am Nat 119:420-426

Eickwort GC, Ginsberg HS (1980) Foraging and mating behavior in Apoidea. Annu Rev Entomol 25:421-466

Fisher RA (1958) The genetical theory of natural selection, 2nd edn. Dover, New York

Hamilton WD (1979) Wingless and fighting males in fig wasps and other insects. In: Blum MS, Blum NA (eds) Sexual selection and reproductive competition in insects. Academic Press, New York, pp 167-220

Holmes HG (1974) Patterns of sperm competition in Nasonia vitri- pennis. Can J Genet Cytol 16:789 795

Houston TF (1970) Discovery of an apparent male soldier caste in a nest of a halictine bee (Hymenoptera: Halictidae) with notes on the nest. Aust J Zool 18:345-351

Huxley J (1938) Darwins theory of sexual selection. Am Nat 72 : 416-433

Jayasingh DB, Taffe CA (1982) The biology of the eumenid mud- wasp Pachyodynerus nasidens in trapnests. Ecol Entomol 7: 283-289

Jolicoeur P (1963 b) The multivariate generalization of the allomet- ric equation. Biometrics 19:497-499

Kinomura K, Yamauchi K (1987) Fighting and mating behaviors of dimorphic males in the ant Cardiocondyla wroughtoni. J Ethol 5:75-81

Klostermeyer EC, Mech SJ Jr, Rasmussen WB (1973) Sex and weight of Megaehile rotundata (Hymenoptera: Megachilidae) progeny associated with provision weights. J Kans Entom Soc 46:536-548

Kukuk PF, Schwarz M (1987) Intranest behavior of the communal sweat bee Lasioglossum ( Chilalictus) erythrurum (Hymenoptera: Halictidae). J Kans Entomol Soc 60 : 58-64

Kukuk PF, Schwarz M (1988) Macrocephalic male bees as func- tional reproductives and probable guards. Pan-Pac Entomol 64:131-137

LaBarbera M (1989) Analyzing body size as a factor in ecology and evolution. Annu Rev Ecol Syst 20:97-117

Metcalf RA (1980) Sex ratios, parent-offspring conflict, and local competition for mates in the social wasps Polistes metrieus and Polistes variatus. Am Nat 116: 642-654

Michener CD, Brothers DJ (1971) A simplified observation nest for burrowing bees. J Kans Entomol Soc 44:236-239

Murray MG (1987) The closed environment of the fig receptable and its influence on male conflict in the Old World fig wasp, Philotrypesis pilosa. Anim Behav 35:488-506

Page RE, Metcalf RA (1982) Multiple mating, sperm utilization, and social evolution. Am Nat 119:263-281

Parker GA (1970) Sperm competition and its evolutionary conse- quences in insects. Biol Rev 45 : 525-567

Plateaux-Qu+nu C (1983) Le volume d'un pain d'abeille influence-t- il le sexe de l'oeuf pondu sur lui? Etude exp6rimentale portant

247

sur la premi6re couve~e d'Evylaeus calceatus (Scop.) (Hymenop- tera, Halictidae). Ann Sci Nat Zool 5:41-52

Plowright RC, Jay SC (1977) On the size determination of bumble bee castes (Hymenoptera: Apidae). Can J Zool 55:1133 1138

Rozen JG Jr (1970) Department of entomology report 102. Annu Pep Am Mus Nat Hist 102:4~5

Rutowski RL, Alcock J (1980) Temporal variation in male copula- tory behavior in the solitary bee Nomadopsispuellae (Hymenop- tera: Andrenidae). Behaviour 73 : 175-188

Shinn AF (1967) A revision of the bee genus Calliopsis and the biology and ecology of C. andreniformis (Hymenoptera: An- drenidae). Univ Kans Sci Bull 46 : 753 936

Snodgrass RE (1956) Anatomy of the honey bee. Comstock, Ithaca London

Sokal RR, Rohlf FJ (1981) Biometry, 2nd edn. Freeman, New York

Starr CK (1984) Sperm competition, kinship and sociality in the aculeate Hymenoptera: In: Smith RL (ed) Sperm competition and the evolution of animal mating systems. Academic Press, Orlando, pp 428 464

Stockhammer KA (1966) Nesting habits and life cycle of a sweat bee, Augochlora pura. J Karts Entomol Soc 39:157-192

Stuart RJ, Francoeur A, Loiselle R (1987a) Fighting males in the ant genus Cardiocondyla. In : Eder J, Rembold H (eds) Chemis- try and biology of social insects (Proceedings of the Tenth Inter- national Congress of the International Union for the Study of Social Insects, Munich, 1986). Peperny, Munich, pp 551-552

Stuart RJ, Francoeur A, Loiselle R (1987b) Lethal fighting among dimorphic males of the ant, Cardiocondyla wroughtoni. Natur- wissenschaften 74:548 549

Timberlake PH (1954) A revisional study of the bees in the genus Perdita F. Smith with special reference to the fauna of the Pacific coast (Hymenoptera: Apoidea) I. Univ Calif Publ Ento- mol 9: 345432

Timberlake PH (1968) A revisional study of the bees in the genus Perdita F. Smith with special reference to the fauna of the Pacific coast (Hymenoptera: Apoidea VII. Univ Calif Publ En- tomol 49:1 196

Wilkes A (1966) Sperm utilization following multiple insemination in the wasp Dahlbominusfuscipennis. Can J Genet Cytol 8 : 451- 461