Original article

Rest at night in some solitary bees - a comparisonwith the sleep-like state of honey bees

W Kaiser

Institut für Zoologie der Technischen Hochschule Darmstadt, Schnittspahnstr, 3,64287 Darmstadt, Germany

(Received 15 November 1994; accepted 26 February 1995)

Summary — The nightly resting behaviour of some solitary bees (Epeolus, Triepeolus, Protepeolus,Anthophora, and Melecta) was studied in the laboratory and compared with results obtained pre-viously in honey bees. Similarities and differences in physiological and behavioural correlates of restwere found; only honey bees exhibit marked amounts of antennal motility.

Apoidea / solitary bee / Apis mellifera / rest at night / sleep / behavioural physiology

INTRODUCTION

"An object in motion always attracts the atten-tion of children, young and old; a butterfly flit-ting from blossom to blossom, a locust jump-ing before one in the dusty road, a beerummaging in a flower, all arouse one’s inter-est. But naturalists, like children, cease topay attention to insects when the latter cease

their activity. Thus the interesting problem ofwhen, where and how insects sleep has beenall but neglected" (Rau and Rau, 1916).

This quotation forms the introduction tothe Raus’ series of lively and succint descrip-tions of field observations of ’sleeping’insects. It also explains why we only acci-dentally discovered the sleep-like state ofhoney bees, during neurophysiologicalinvestigations of their visual system (Kaiserand Steiner-Kaiser, 1983).

Rau and Rau used the term ’sleep’ intu-itively although they had also performedsome simple experiments. They had dis-covered, for instance, that it became slowlybut increasingly difficult to disturb animalswhich had come to rest. In addition, ’sleep-ers’ required a certain amount of time to’wake up’ in the morning. One further impor-tant observation on solitary wasps and soli-tary bees was that marked individuals ofcertain species repeatedly returned to par-ticular sites to ’sleep’ in the company of con-specifics.

Since the term ’sleep’ implies statementsabout the level of consciousness, it can onlybe strictly applied to humans. However, theterm can justifiably be used for animalswhose rest exhibits those objectively mea-surable phenomena which accompanyhuman sleep (’sleep signs’). We therefore

used a catalogue of criteria provided byTobler (1984) when we studied nightly restin the honey bee. We employed bothelectrophysiological techniques (elec-tromyogram-recordings) and methods usedin quantitative behavioural physiology. Theresults clearly demonstrate the existenceof a sleep-like state in this insect (Kaiser,1988). The studies reported here wereprompted by the following question: is thesleep-like state in the honey bee, with itsmany parallels to mammalian sleep, a sin-gular phenomenon restricted to this highlyevolved social bee? A comparison betweenthe nightly rest in honey bees and that ofnon-social, solitary bees should yield a cata-logue of similarities or differences and thusan answer to this question.

This paper is the first detailed account

of laboratory investigations on the nightlyrest of solitary bees. The solitary bees whichwere studied normally spend the night eitherin the open or in earther burrows. The former

display a bizarre resting behaviour: theyattach themselves to the end of a twig onlywith their mandibles. Numerous field obser-

vations of this behaviour have been pub-lished (Fiebrig, 1912; Rau and Rau, 1916;Rau, 1938; Schremmer, 1955, 1961;Westrich, 1989). Preliminary reports of ourwork have appeared previously (Kaiser,1990a, b; Kaiser and Steiner-Kaiser, 1991).

MATERIAL AND METHODS

Experimental animals

Epeolus variegatus (Linnaeus 1758,Anthophoridae). Experimentsin Portugal in June 1989

These brood parasites were caught during thelate afternoon near Costa da Caparica while feed-ing on Chrysanthemum coronarium, a yellow

composite. The animals were brought back tothe laboratory in small glass bottles containingsoft tissue paper which were kept in the dark. Inthe dimly lit laboratory, the bees were transferredto a cage which contained small dry twigs setvertically into a piece of plasticine. Initially, thebees walked and flew around the cage. After a

while they became increasingly quiet and walkedup and down the twigs. Finally, each bee clampedon to the tip of a twig with its mandibles (headpointing downwards). If they were disturbed soonafter clamping on, the bees left the twig. I there-fore waited 15 min before transferring an Epeolusplus twig to the experimental chamber. Each ani-mal was used only once for an experiment.

Epeolus sp (Anthophoridae)

An Epeolus ♀ (species unknown) was captured ona flower in Douglas, Arizona on September 13 1989 and studied for one night.

Triepeolus sp (Anthophoridae). Experi-ments in Arizona, USA, during Augustand September 1989

These brood parasites were also captured whilethey were feeding. They were found on yellowcomposites (Verbesina sp) near Rodeo, NewMexico. Due to the very high ambient tempera-tures, the bees were brought back to the labora-tory in cooled, light-tight containers.

The specimens of Triepeolus could be used forseveral experiments on different nights, if theywere maintained in outdoor cages containing avase with a small bunch of fresh Verbesina and

Sphaeralcea (mallows). The cages were exposedto direct sunlight, but were moved at irregularintervals to shady areas, to avoid overheating theoccupants. One bee (V8) survived under theseconditions from September 9 to 24.

All the Triepeolus specimens could find theirresting sites in complete darkness. In the earlyevening, they were taken out of their cages andtransferred individually to glass beakers whichcontained small vertically oriented twigs. Thebeakers were placed in complete darkness. Usu-ally, the bees had clamped on to the tip of a twig(head pointing downwards) within 5 min.

Protepeolus singularis (Anthophoridae)

One female and 1 male specimen of this broodparasite were caught near Rodeo, NM on August24 1989 and August 27 1989 respectively. Theywere handled in the same way as Triepeolus.

All these solitary bees (Epeolus, Triepeolusand Protepeolus) normally spend the night in theopen, clamped onto a twig with their mandibles.

Anthophora fulvitarsis (Brullé) andMelecta aff tuberculata (Lieftinck) (bothAnthophoridae). Portugal, June 1989

M tuberculata is a brood parasite (cuckoo) of A ful-vitarsis. Females of both species were capturedin the evening at a loess cliff. The cliff faces east-wards and was covered with hundreds ofentrances to the nest burrows of Anthophora.Both the solitary bee Anthophora and its cuckoospend the night in the burrows, which Anthophoradigs. The site is located near Alfarim, southeast ofLisbon and around 4 km from the Atlantic coast.The animals were brought to the laboratory (seeEpeolus, above) and each was studied for onenight.

Experimental apparatus

The behaviour of the bees was observed in totaldarkness with the aid of an infrared-sensitivevideo camera (Panasonic CD22) connected to atimelapse recorder (Panasonic NV-8050). Fourlight-emitting diodes (LD 271, maximum emis-sion at 950 nm) served as the infrared source.Two pieces of opaque acrylic glass, locatedbetween the diodes and the animal, provided auniformly illuminated background. The camerathus viewed the bee in silhouette in front of a lightbackground.

In the case of Epeolus, Triepeolus and Pro-tepeolus, the twig, together with its attached bee,was mounted in front of the opaque plates. Theoriginal spatial orientation of the twig in the cagewas usually preserved. Exceptions have beennoted in the legends of the appropriate figures.

Individual specimens of Melecta and

Anthophora were placed in a rectangular trans-parent glass-walled cuvette for observation atnight. The narrow sides and the floor of the cuvettewere made of Plexiglas and covered with thin

balsa wood; the lid, also of Plexiglas, had holes forventilation. The space available to an animal insidethe cuvette measured 44 x 33 x 16 mm.

Data analysis

All the video tapes were evaluated on a TV-mon-itor during playback. When greater temporal re-solution was required, the tapes were played backat the original recording speed (the timelapse fac-tor during continuous recording was either 4 or 12)and measurements were made on single frames.

RESULTS

Epeolus, Triepeolus and Protepeolus

Laboratory experiments

Under moderate room-light conditions, theclamping-on behaviour could be observeddirectly. The bees crawled up the twigs,went over the top and then clasped the twig,at or near its tip, with their mandibles (fig1). Usually, they then raised their legs fromthe substrate and began to intensivelygroom their bodies (fig 2b). Grooming wasfollowed by an observable change in thebees’ behaviour: if the animals were dis-

turbed soon after they had clamped on toa twig, they tended to release their hold andstarted moving down the twig. This tendencywas much weaker after a grooming session.Soon after the bees had clamped on, andafter grooming, the antennae pointed for-ward and lay very close to, or even touched,the twig (figs 1, 2a). The antennae ofTriepeolus and Protepeolus then movedvery slowly in an arc towards the animals’dorsal surface (fig 2c, d). Their final restingposition was independent of the spatial ori-entation of the animals and of gravity: thesame antennal position relative to the headas in figure 2d was reached by animals,

which clamped on to vertical twigs with theirheads pointing downwards.

Epeolus also touched the twig with itsantennae or held them close to it in the

evening, after clamping on, and in the morn-ing, just prior to releasing its mandibles.During the night, however, the antennaemoved to a dorsolateral position. After theyhad reached their final resting position, theyremained motionless. This was also true for

Triepeolus and Protepeolus. However, allanimals made antennal movements duringgrooming and other large body movements.

The position of the head remainedunchanged throughout the night, indicatingthat the mandibular muscles maintainedtheir tone. The thorax and abdomen, on theother hand, slowly adopted different posi-

tions, indicating a decrease of tone in othermuscles (compare fig 2a with 2d).

The upper curve in figure 3 demonstratesthe time course of changes in the spatialorientation of the thorax of an E variegatus(P11). The position of the thorax was mea-sured with a protractor attached to thescreen of the monitor. The bee had clampedon to the end of a twig in a spatial orientationapproximately corresponding to that in figure1 c. The angle a was defined as follows: theangle between a vertical reference line(direction of gravity) passing through themidpoint of the bee’s neck and the line con-necting this point with the scutellum (dorsallandmark on the thorax; compare also withthe inset of fig 6 below, where the referenceline is horizontal).

Further movements that bees performedafter having clamped onto a twig werepumping abdominal movements. Theseoccurred in bursts which followed one

another at more or less long intervals. Thelower curve in figure 3 represents the timecourse of the duration of ventilatory cycles(bursts plus intervals) in the bee P11. Thenumber of pumping movements per burst,and the duration of individual bursts,remained relatively constant for long peri-ods during the experiment. For the intervalbetween 4.15 and 4.30 h, there were, on

average, 85.5 pumping movements perburst and the average burst duration was27.2 s. After approx 6.30 h, the number of

pumping movements per burst decreasedand the burst duration increased.

Each burst of ventilatory movements waspreceded by a succession of abdominal

extensions. These abdominal extensionswere accompanied by an elevation of thethorax and abdomen, when the animalsrested in postures like those shown in fig-ures 1 and 2. After the end of each series ofabdominal pumping movements, the tho-rax and abdomen slowly sank down again.The time course of the size of the angle a infigures 3, 5B and 6 (below) does not showthe dynamic changes induced by ventila-tion, since a was always measured at theend of a ventilatory cycle, before theabdomen stretched. Figure 3 shows thatventilation became slower as the bee’s bodysank downwards. There was a spontaneousincrease in ventilatory frequency around4.30 h.

Of the solitary bees I studied in the lab-

oratory, the animal which yielded the resultsfor figure 3 showed the least motility during

the night. During an interval of 14.6 h, it dis-

played neither large body movements norgrooming, although the ambient tempera-ture (24.5°C) was higher than under naturalconditions. The longest episode of rest(absence of larger body movements andgrooming) of the bee in figure 4 lasted 3 hand occurred late at night. Figure 4 illus-trates a further typical result: episodes ofgrooming (of head, body and antennae)were longest during the transition from activ-ity to rest, and most frequent during the tran-sition from rest to activity. The longestgrooming episode in figure 4 lasted

23.4 min!

The forms of the curves in figure 4 aretypical for the results obtained from all beesthat clamped on; however, there were dif-ferences in the absolute values. The majorresults from these experiments are sum-marized in table I.

The experiment shown in figure 5 wasprompted by the question whether clamp-ing-on is a necessary precondition for reach-ing a stable state of rest. In the early eveningof September 15 1989, a Triepeolus, V8,was placed in a transparent cuvette linedwith balsa wood (inner dimensions 22 x 16 x 6 mm). As figure 5A shows, the animal didnot enter a prolonged state of rest but dis-

played instead a variety of behaviour (loco-motion, grooming and rest). I terminated theexperiment at 22.06 h. The room lightingwas switched on and the bee was trans-ferred to a beaker containing a vertically ori-ented twig. At first, the bee walked animat-edly around in the beaker, but, within 7 min,it had clamped on to the tip of the twig. Sub-sequently, neither carrying the twig around inthe illuminated laboratory, nor mounting it

into the experimental apparatus, promptedthe bee to release its hold. The animal’s

behaviour during the rest of the night isshown in figure 6. These subsequent resultswere again obtained in total darkness. Tocompare the bee’s behaviour in the cuvette

(fig 5A) with the behaviour one could expectit to show at the same time of day, but whileclamped on to a twig, the data from the firsthalf of the next night are presented in figure5B. There are 3 major differences betweenfigures 5A and 5B: (1) the bee only walkedaround in the cuvette; it never released itshold on the twig; (2) long-lasting groomingepisodes occurred repeatedly in the cuvette;on the twig, the bee groomed intensivelyonly during the first hour of the experimentand subsequently only briefly and sporadi-cally; and (3) between approximately 21 and22 h, rest in the cuvette was almost negligi-ble, whereas it occupied the entire interval onthe twig.

’Field experiments’

The data presented above were obtainedunder constant laboratory conditions. Thefollowing data were derived from 2 ’fieldexperiments’. An optical bench which carriedthe video camera, the infrared light sourceand the holder for the twig plus bee was setup under the open sky in the garden out-side the laboratory at each location. The

experimental animals were thus exposedto the naturally-occurring variations in lightintensity, air temperature, and humidity.

Epeolus sp (P17), Portugal,June 26-27, 1989

The bee clamped on to the tip of a verti-cally-oriented twig (head pointing down-wards) at 19.45 h; the ventral side of the

animal’s thorax rested on the end of the twig.The original orientation of the twig was main-tained when it was mounted on the opticalbench. Air temperature decreased progres-sively during the night and reached its min-imum of 15.8°C at 5.05 h. This valueremained constant till sunrise, at 6.30 h. Inthe morning, around 10.00 h, the weatherwas clear and sunny, and the air temperaturehad reached 23.3°C. The bee, still clampedon, was in the shade of the laboratory build-ing. It displayed occasional body movementsand breathed more quickly than during thenight (the duration of one ventilatory cyclewas 0.86 min as compared to 9.1 min at theminimum temperature). A few minutes later,the bee was in direct sunlight and, a fewminutes after that, it flew away. The evalua-tion of the data yielded curves very similar to

those in figure 4; only the absolute valuesdiffered. The longest rest episode lasted 6.6h, the longest ventilatory cycle 9.1 min. Thedescending flanks of the curves were shiftedtowards earlier times of day.

One event, which occurred late at night, isworth mentioning. The resting bee was vis-ited by an ant. As the ant crawled up thetwig, it encountered the bee’s head. The beereacted by making very pronounced defen-sive leg movements and the ant movedaway and walked down the twig soon after.

Triepeolus sp (V8), Arizona,September 23-24, 1989

The resting behaviour of this animal hadalready been studied in the laboratory; theexperiment was conducted similarly to the

one just described. However, the spatial ori-entation of twig and animal was different.After the animal had clamped onto the twigwhen this was in a vertical position, the twigwas mounted horizontally in the apparatusso that the dorsal surface of the animalfaced the ground (see fig 2 and inset to fig7). In addition, in Arizona, the nightly mini-mum temperature (4.9°C) was much lowerthan in Portugal and also lower than thenightly minimum at the valley location wherethe bee had been captured (10.4°C). Thislatter difference is due to the high altitudeof the Research Station.

The low ambient temperature led to asurprising result: the animal showed no dis-continuous resting ventilation after the tem-perature dropped below approximately 9°C,and it only resumed breathing in this fashionin the morning when the temperature rosebeyond 10°C (fig 7). The bee’s antennaealso assumed an unexpected resting posi-tion during the night. Instead of adoptingthe position typical of long rest episodes inthe laboratory (fig 2d), the antennaeremained in the position shown in the insetin figure 7 during most of the night. Thisresult was unexpected in view of the fact

that the duration of the bee’s longest restepisode (14.5 h) was much greater than thatfound under laboratory conditions in thesame animal (see table I). A new phe-nomenon appeared at low temperatures:leg twitches. Like Epeolus in Portugal,Triepeolus released its hold on the twig onlyafter it became exposed to direct sunlight(ambient temperature approx 30°C).

A fulvitarsis, M tuberculata

Observations at nest sites of A fulvitarsis

Both A fulvitarsis and its brood parasite, Mtuberculata, spend the night in burrows dugby A fulvitarsis. On the observation days(June 8, 10 and 14, 1989), the weather wasclear and sunny; only females wereobserved.

Evening observations. Several A fulvi-tarsis returned from foraging flights aftersunset (eg, return time 21.06 h, sunset20.50 h). The temperature in the burrowsafter sunset was higher than the ambienttemperature outside (burrow: 19.9°C; airtemperature: 14.8°C). The animals restedwith their heads directed towards the bur-row entrance (observation in 3 bees).

Morning observations. A few specimensof A fulvitarsis were seen returning to theirburrows at 6.13 h, well before sunrise (6.35h). The air temperature at this time was12.7°C, while the temperature in the bur-rows was 17.4°C. Several A fulvitarsis couldalso be seen sitting at their nest entrancesbefore sunrise, but they only flew off afterthey had sat in direct sunlight for a whileand had groomed themselves extensively.

Flying M tuberculata were first encounteredat 7.40 h, when the air temperature was17.4°C.

Laboratory experimentsFigure 8 represents the results obtained byobserving a female specimen of A fulvitarsis(P1) which spent the night in a cuvette at anambient temperature of 22.3°C. The animalrested on the floor of the cuvette. The longestrest episode (interval without head, body, orgrooming movements) occurred late in thenight, and had a duration of 1.3 h. The entireexperiment was performed in complete dark-ness, but the bee became active at 6.17 h,before sunrise (6.30 h). The animal displayeddiscontinuous resting ventilation at night,with the longest ventilatory cycle also occur-ring late at night. The stepwise abdominalelongation before the start of ventilatorymovements, displayed by Epeolus, Prote-peolus, and Triepeolus, also occurred in Afulvitarsis. Grooming movements wereabsent over a long interval late at night.

Two phenomena which accompanied thenightly rest of this bee are not shown in fig-ure 8: a decrease in leg muscle tone (theanimal gradually sank onto the floor of thecuvette during longer rest episodes) andtarsal twitching. In addition, during long restepisodes, the antennae adopted character-istic postures: after being initially held parallelto the floor of the cuvette, they graduallysank down until the flagella were approxi-mately parallel to the frons of the animal.No further antennal movements wereobserved until a bout of grooming or a largebody movement occurred. The sameapplied to Melecta (see below).

Observations on 2 further female A ful-vitarsis (P2 and P8) yielded results very sim-ilar to those in figure 8, but rest was morepronounced: longest rest episode: 3.0 h;longest average ventilatory cycle: 16.6 min(see table I). Both animals became spon-taneously active early in the morning, butonly after sunrise (see table I). Two femalespecimens of M tuberculata, which where

observed individually on 2 nights, also incomplete darkness, behaved very much liketheir hosts: they also rested on the floor ofthe cuvette (ventral side down), exhibitedlong episodes of rest which were accom-panied by slower breathing and reductionin leg muscle tone, as well as spontaneousearly return to activity. However, the 2species differed clearly with respect toantennal posture: the antennae of M tuber-culata moved slowly backwards duringlonger rest episodes and finally came to liein a dorsolateral position. Further details ofthese experiments are provided in table I.

DISCUSSION

As long ago as 1916, Rau and Rau sug-gested investigating the physiological pro-cesses accompanying ’sleep’ in insects. Todate, very few scientists have responded.The most recent studies are on cockroaches

(Tobler, 1983; Tobler and Neuner-Jehle,1992) and honey bees (see below).

The work reported here was initiated byour interest in the question of whether thephysiological processes during the nightlyrest of the honey bee (Kaiser, 1988) differfrom those of related, less highly developed(solitary) bees. In this paper nightly rest insolitary bees is studied for the first time usingmethods of quantitative behavioural physio-logy. A valid basis for comparison is pro-vided by the fact that most of the experi-ments with solitary bees (like those on thehoney bee) were performed under constantlaboratory conditions and at a temperaturehigher than 20°C. Both similarities and dif-ferences have emerged.

Similarities

Resting sites

Both honey bees and solitary bees seek outspecific places at which they spend the

night. Pollen foragers of the honey bee (onlythis type of worker bee has been studiedso far) rest in the hive, usually either on thecombs, outside the brood area, or on thehive walls (Sauer and Kaiser, 1995). Thefact that these animals do not participate inthe generation of heat in the hive showsthat they are truly resting (Kaiser, 1988).

The biological significance of the fact thatsome solitary bees clamp on to the end of atwig may be increased protection againstenemies. In this position, a bee could bemistaken for part of the plant, and it would

also be less accessible to ants (seeResults). Both hives and nest burrows offerhoney bees and A fulvitarsis, together withits brood parasite M tuberculata, respec-tively, increased protection at night.

Phenomena accompanying rest at night

In both honey bees (Kaiser, 1988) and thespecies of solitary bees investigated here,observations under constant laboratory con-ditions revealed, at night, the presence ofprocesses which have similar time courses.

(1) There was a progressive increase in theduration of rest episodes, with maximumvalues occurring late at night. (2) The ven-tilatory frequency fell to a minimum late atnight (honey bees; unpublished data). (3)The tone in certain muscles decreased. (Inthe solitary bees, changes in the orienta-tion of the whole, or parts, of the body wereinterpreted as decreases in muscle tone, if

these changes occurred gradually and inthe direction of gravity. In honey bees, acomparison between gradual changes inhead position and the amplitude of neck-muscle myograms proved that there was aprogressive loss of tone.) (4) The animalsadopted characteristic antennal postures.

The following phenomena also accom-panied rest at night in all species: twitchingmovements of the tarsi, and increased

grooming during the transitions from activityto rest and rest to activity. The phenomena

(1) to (4) indicate that solitary bees exhibit aprogressive increase in the depth of rest atnight. This conclusion is based upon a com-parison of the present results with thoseobtained in the honey bee. In this animal,the time course of the nightly reactionthreshold (fig 9 in Kaiser, 1988) agrees wellwith the time courses of the 4 parametersdescribed above.

Differences

Motility during rest at night

Even at 25°C, a temperature which, at night,is normally available only to honey bees,solitary bees exhibited much less motilitythan their social relatives. The longest restepisode (absence of head and body move-ments) measured to date in honey bees at25°C is 13.4 min (Kaiser, 1988). At almostthe same temperature (24.5°C), the longestrest episode measured in A fulvitarsis (P8)was 1.5 h. At this temperature, E variegatus(P11) displayed a rest episode whose dura-tion was 14.6 h! Motility during the nightclearly differs between honey bees and thesolitary bees I have studied. This becomeseven more obvious when one comparesthe amount of antennal immobility. Thelongest episode of antennal immobility mea-sured to date in honey bees is 4.1 min(Kaiser, 1988). In the solitary bees, theintervals without antennal movements

lasted for hours late at night. It is likely that,under natural conditions, honey bees andthe solitary bees which spend the nightcompletely in the open differ even moremarkedly with respect to their motility dur-ing rest. This assumption is supported bythe results of the ’field experiments’ on Evariegatus and Triepeolus sp (V8). A fulvi-tarsis has not been investigated experi-mentally under near-natural conditions, todate. This species, under natural conditionsat night, probably does not cool down to

the same extent as Epeolus, Triepeolus,and Protepeolus, for the following reasons:(1) A fulvitarsis spends the night in earthenburrows which are warmer than the air out-

side; (2) it is considerably larger than theother 3 species; (3) it has thick ’fur’. A ful-

vitarsis is a heterothermic insect (seeResults; observations at nest sites). Thisspecies may therefore even actively main-tain its body temperature at a particularvalue during the night. Stone (1993) hasdemonstrated, for A plumipes, that highbody mass is advantageous in the genera-tion of higher thoracic temperatures whenthe ambient temperature is low.

A fulvitarsis might therefore displayapproximately the same motility at nightunder natural conditions as it did in the lab-

oratory at a temperature of 21-25°C. Buteven at these temperatures, this solitarybee displayed much less motility at rest thanthe honey bee (table I). M tuberculata, thebrood parasite of A fulvitarsis, is about thesame size as its host and the factors men-

tioned for its host probably are relevant to it,too.

The relationship between resting timeand light intensity

Honey bees continue to forage after sun-set and they become active in a dark hive inthe morning. When kept in continuous dark-ness in the laboratory, they display a circa-dian (endogenous) rest-activity rhythm (seeKaiser, 1988, for review). The solitary beeswhich clamp on to a twig to rest behavequite differently. They can adopt this pos-ture when the weather starts becomingcloudy (Westrich, 1989, p 124, 267, Kaiser,unpublished data - Triepeolus sp). In the 2’field experiments’, both E variegatus andTriepeolus sp became active only after theyhad been exposed to direct sunlight, despitethe fact that the ambient temperature in theshade was more than 20°C. It thus seems

likely that, under natural conditions, theseanimals do not become active in the dark. In

the laboratory, the animals were observed incontinuous darkness and at unnaturally hightemperatures. Under these conditions, theybecame active (released their hold on thetwig) either late in the morning (eg, P11, fig3) or in the early afternoon (eg, P12, fig 4);table I provides further examples. Theresumption of activity might have been trig-gered either by the depletion of their energyreserves or by an internal clock. Thesehypotheses are presently under investiga-tion.

In contrast to the bees which clamp on, Afulvitarsis, like honey bees, can fly after sun-set and they also become active in their bur-rows before sunrise. In the laboratory, inconstant darkness, they also resumed activ-ity early in the morning (fig 8, table I).

Ambient temperature at night

Honey bees spend the night in the hive,where temperatures around 25°C are com-monly encountered outside of the broodarea (Hess, 1926). This temperature is wellabove 12°C, the ’chill coma’ limit in honeybees (Lighton and Lovegrove, 1990, Apismellifera ligustica). Below an ambient tem-perature of 12°C, honey bees shift from dis-continuous, convective ventilation to con-

tinuous, diffusive ventilation. In the

experiment with Triepeolus sp (V8) undernear-natural conditions, discontinuous rest-

ing ventilation stopped when the tempera-ture fell below 9°C and resumed when the

morning temperature reached approximately10°C again (see fig 7). The nightly temper-atures in the areas where these bees occur

naturally can reach such values. The tem-perature in the nest burrows of Anthophoraremains higher than the ambient air tem-perature and thus very likely above the ’chillcoma’ limit. The temperature at which A ful-vitarsis enters ’chill coma’ is unknown.

Clamping-on behaviour

This bizarre posture that Epeolus, Protepe-olus and Triepeolus exhibit while resting isdramatically different from the resting pos-tures of Anthophora, Melecta, and honeybees. Clamping-on is an important prereq-uisite for attaining a state of profound rest.Clamping-on behaviour during rest alsooccurs in other Hymenoptera, including beeswhich are not brood parasites.

CONCLUSIONS

Rest at night in all the bee genera studied todate is characterized by a number of com-mon features which all show that none of

these insects becomes rigid during thisphase.

However, one conspicuous differencehas emerged: only honey bees exhibit largeamounts of motility during rest. In particu-lar, their antennae are much more motilethan those of the solitary bees. This indi-cates that active processes occur duringthe sleep-like state of honey bees. Sleep inmammals is also characterized by activeprocesses (Borbély, 1986; Koella, 1988).We can, at present, only speculate aboutthe biological significance of processes dur-ing rest at night in the honey bee. Theycould be basically related to the complexlife style of these social insects. The highambient temperature in the honey bee hiveat night might be a necessary preconditionwhich allows complex physiological pro-cesses to occur (eg, alternation betweenantennal movements and antennal immo-

bility; changes in muscle tone, reactionthreshold, and neuronal activity).

Those solitary bees which clamp on totwigs at night rest in a very different waythan honey bees. The profound rest exhib-ited by these solitary bees could be a tor-por-like condition. Conservation of energyis probably much more important for these

animals than for honey bees. The type ofnightly rest shown by those species of soli-tary bees which spend the night in burrowscould occupy a position somewherebetween these 2 extremes.

ACKNOWLEDGMENTS

I thank the Deutsche Forschungsgemeinschaft(Program SFB 45) and the late O Schulz-Kampfhenkel (founder of the biological researchstation Quinta de Sao Pedro, Portugal) for gen-erous financial support. A Pircher and P Westrichprovided invaluable help in Portugal, as did WSherbrooke, Director of the American Museum ofNatural History’s Southwestern Research Sta-tion, Portal, and J Rozen and H Spangler in Ari-zona. G Bayer patiently prepared the figures andtyped the manuscript. I thank her and the numer-ous students who participated in the data evalu-ation. H Niemetz constructed some of the exper-imental apparatus. My wife, Jana Steiner-Kaiser,translated the manuscript, and contributed criti-cal discussions and practical support through-out this project. Two anonymous referees madevaluable suggestions for improving the

manuscript.

Résumé — Repos nocturne chez

quelques abeilles solitaires (Hymenop-tera, Apoidea). Comparaison avec l’étatproche du sommeil chez l’abeille domes-tique (Apis mellifera L). Les abeilles soli-taires étudiées passent la nuit dans lanature, agrippées par les mandibules à desplantes sèches (Epeolus, Tripeolus, Prote-peolus, voir fig 1) ou dans des nids creu-sés dans des falaises de loess (Anthophorafulvitarsis et le parasite de son couvainMelecta tuberculata). Les insectes ont étécapturés sur des fleurs ou à l’entrée desnids. Leur comportement a été ensuite enre-

gistré avec une caméra vidéo sensible àl’infrarouge en laboratoire, en conditionsconstantes et à l’obscurité totale. La com-

paraison des résultats avec ceux obtenuschez l’abeille domestique (Kaiser, 1988)montre des similitudes aussi bien que des

différences. Similitudes : les abeilles domes-

tiques comme les abeilles solitaires possè-dent des lieux spécifiques de repos. Pen-dant la nuit la durée des périodes de repos(absence de toilettage et de mouvementscorporels amples) augmente (figs 4, 5, 6 et8), alors que le rythme respiratoire diminue(figs 3, 4, 7 et 8 ; tableau I) et que la ten-sion de certains muscles est réduite (figs 2,3, 5 et 6). Quand le repos se prolonge, lesantennes prennent une position caractéris-tique (figs 2, 5B, 6 et 7). Ces phénomènesindiquent un approfondissement progressifdu repos au cours de la nuit, car ils s’accom-pagnent chez l’abeille domestique d’uneélévation du seuil de réaction. Différences:la motilité du corps et des antennes au coursdu repos nocturne est beaucoup plus réduitechez les abeilles solitaires que chez lesabeilles domestiques. Les espèces quis’accrochent par les mandibules sont parti-culièrement calmes sans être immobiles.La durée du repos dépend beaucoup desfacteurs externes, en particulier de lalumière. Tandis que les abeilles domes-

tiques et Anthophora sont déjà actives avantle lever du soleil ou encore après le cou-cher, un individu Epeolus étudié en condi-tions naturelles, ainsi qu’un individu Trie-peolus (fig 7) n’ont quitté leur emplacementde repos que lorsque les rayons du soleill’ont touché directement. L’abeille domes-

tique se distingue par le fait qu’elle se pro-tège du «coma» dû au froid en passant lanuit dans la ruche chaude. La températureélevée de la ruche pourrait être la conditionnécessaire pour qu’aient lieu, durant l’étatproche du sommeil, les processus com-plexes qui s’expriment avant tout par l’alter-nance du repos et du mouvement desantennes et par les modifications du tonus

musculaire, du seuil de réaction et de l’acti-vité neuronale. Le sommeil des mammi-fères se caractérise lui aussi par des pro-cessus actifs. Chez les abeilles solitaires

qui passent la nuit dehors et s’agrippent parles mandibules, la nécessité d’économiserl’énergie contraint à un état proche de la

torpeur. Le type de repos nocturne desabeilles qui passent la nuit dans des trouspourrait représenter un intermédiaire entreles 2 extrêmes.

Apoidea / abeille solitaire / Apis melli-fera / sommeil / repos nocturne / phy-siologie du comportement

Zusammenfassung — Die Nachtruheeiniger Solitärbienen: ein Vergleich mitdem schlafähnlichen Zustand von

Honigbienen. Die untersuchten Solitär-bienen übernachten in der Natur festge-bissen an trockenen Pflanzen (Epeolus,Triepeolus, Protepeolus, siehe auch Abb1) oder in Bruthöhlen in Lößwänden (Ant-hophora, die die Höhlen gräbt, und ihr Brut-parasit Melecta). Die Tiere wurden an Blü-ten oder vor den Bruthöhlen gefangen; ihrVerhalten wurde dann im Labor unter kon-stanten Bedingungen und bei völliger Dun-kelheit mit einer infrarotempfindlichenVideoanlage registriert. Die Ergebnissewurden mit den Befunden von Honigbie-nen (Kaiser, 1988) verglichen und sowohlÄhnlichkeiten als auch Unterschiede gefun-den. Ähnlichkeiten: Sowohl Honigbienenals auch die Solitärbienen haben spezifi-sche Ruheorte. Während der Nacht nimmt

die Dauer der Ruhe-Episoden (Abwesen-heit von Putzen und größeren Körperbe-wegungen) zu (Abb 4, 5, 6, 8), während dieAtmungshäufigkeit abnimmt (Abb 3, 4, 7,8; Tabelle I) und die Spannung bestimm-ter Muskeln sich verringert (Abb 2, 3, 5, 6).Bei längerer Ruhe nehmen die Antenneneine charakteristische Haltung ein (Abb 2,5B, 6, 7). Diese Phänomene deuten aufeine allmähliche Vertiefung der Ruhe imVerlauf der Nacht hin, denn bei Honigbie-nen werden sie von Erhöhungen der Reak-tionsschwelle begleitet. Unterschiede: DieBewegungshäufigkeit des Körpers und derAntennen ist während der nächtlichen Ruheder untersuchten Solitärbienen viel gerin-ger als bei Honigbienen. Die sich fest-

beißenden Arten sind besonders ruhig, abernicht starr. Bei diesen ist die Dauer derRuhe sehr stark von exogenen Faktoren

abhängig, vor allem von Licht. WährendHonigbienen und Anthophora schon vorSonnenaufgang und noch nach Sonnen-untergang fliegen, verließen ein im Freienuntersuchter Epeolus sowie ein Triepeolus(Abb 7) ihren Ruheplatz erst, als sie vondirektem Sonnenlicht getroffen wurden. DieHonigbiene zeichnet sich dadurch aus, daßsie nachts im warmen Stock vor Kältestarre

geschützt ist. Die hohe Stocktemperaturkönnte die Voraussetzung für die komple-xen Vorgänge während des schlafähnli-chen Zustands dieser Tiere sein, die sichvor allem in dem immer wiederkehrenden

Wechsel zwischen Antennenruhe und

Antennenbewegung und in Änderungenvon Muskeltonus, Reaktionsschwelle undneuronaler Aktivität äußern. Auch der Schlafder Säuger ist durch aktive Prozessegekennzeichnet. Bei den im Freien über-nachtenden, sich festbeißenden Solitär-bienen könnte die Notwendigkeit, Energieeinzusparen, den torpor-ähnlichen Zustanderzwingen. Die Art der nächtlichen Ruheder in Höhlen übernachtenden Bienenkönnte eine Mittelstellung zwischen diesenbeiden Extremen darstellen.

Apoidea / Solitärbiene / Apis mellifera /Nachtruhe / Schlaf / Verhaltensphysio-logie

REFERENCES

Borbély AA (1986) Secrets of Sleep. Basic Books, NewYork, USA

Fiebrig K (1912) Schlafende Insekten. Jena Z Natur-wiss 48 (N F41), 315-364

Hess WR (1926) Die Temperaturregulierung im Bienen-volk. Z Vergl Physiol 4, 465-487

Kaiser W (1988) Busy bees need rest, too. Behaviouraland electromyographical sleep signs in honeybees.J Comp Physiol A 163, 565-584

Kaiser W (1990a) Wie solitäre Bienen die Nacht ver-bringen. Verh Dtsch Zool Ges 83, 616-617

Kaiser W (1990b) The behaviour of solitary bees at night.In: Brain-Perception-Cognition. Proc 18th GöttingenNeurobiology Conference (N Elsner, G Roth, eds),Thieme, Stuttgart, Germany, USA, 64

Kaiser W, Steiner-Kaiser J (1983) Neuronal correlates ofsleep, wakefulness and arousal in a diurnal insect.Nature (Lond) 301, 707-709

Kaiser W, Steiner-Kaiser J (1991) Sleep in insects - Acomparison of the nightly resting state in social andsolitary bees. Sleep Res 20A, 223

Koella WP (1988) Die Physiologie des Schlafes. EineEinführung. Fischer, Stuttgart, Germany

Lighton JRB, Lovegrove BG (1990) A temperature-induced switch from diffusive to convective ventila-tion in the honeybee. J Exp Biol 154, 509-516

Rau P (1938) Additional observations on the sleep ofinsects. Ann Entomol Soc Am 31, 540-556

Rau P, Rau N (1916) The sleep of insects; an ecologicalstudy. Ann Entomol Soc Am 9, 227-274

Sauer S, Kaiser W (1995) Pollen foragers of the honeybee (Apis mellifera carnica L) rest in the hive at night.In: Learning and Memory. Proc 23rd Göttingen Neu-robiology Conference (N Elsner, R Menzel, eds),Thieme, Stuttgart, Germany, 253

Schremmer F (1955) Beobachtungen über die Nachtruhebei Hymenopteren, insbesondere die Männchen-schlafgesellschaften von Halictus. Österr Zool Z6,70-89

Schremmer F (1961) Aus dem Nachtleben der Insek-ten. Der Mittelschullehrer und die Mittelschule, 3/4

Stone GN (1993) Endothermy in the solitary beeAnthophora plumipes: independent measures ofthermoregulatory ability, costs of warm-up and therole of body size. J Exp Biol 174, 299-320

Tobler I (1983) Effect of forced locomotion on the rest-activity cycle of the cockroach. Behav Brain Res 8,351-360

Tobler I (1984) Evolution of the sleep process: a phylo-genetic approach. In: Sleep Mechanisms (AA Bor-bély, JL Valatx , eds) (Exp Brain Res, Suppl vol 8),Springer, Berlin, Germany, 207-226

Tobler I, Neuner-Jehle M (1992) 24-h variation of vigi-lance in the cockroach Blaberus giganteus. J SleepRes 1, 231-239

Westrich P (1989) Die Wildbienen Baden-Württembergs,2 Bd, Ulmer, Stuttgart, Germany