Download pdf - ADAPTATIONS AND RESPONSES OF DASYMUTILLA OCCIDENTALIS (HYMENOPTERA: MUTILLIDAE) TO PREDATORS

Ent. exp. & appl. 21 (1977) 99--11 I. North-Holland Publ. Co. Amsterdam

ADAPTATIONS AND RESPONSES OF DASYMUTILLA OCCIDENTALIS (HYMENOPTERA: MUTILLIDAE) TO

PREDATORS

BY

JUSTIN O. SCHMIDT and MURRAY S. BLUM

Department of Entomology, University of Georgia, Athens, Georgia, U.S.A.

The mutillid wasp Dasymutilla occidentalis possesses several adaptations and exhibits a number of responses which appear to be of defensive value: a long mobile sting with powerful venom; a strong, rounded and slippery cuticle; an ability to run very rapidly and evasively; an aposematic warning coloration pattern; and the ability to respond to an attack by making stridulatory sounds and by releasing a chemical secretion or both. The effectiveness of these .defenses is supported by tests utilizing various vertebrate and arthropod predators. The raison d'&re of the multiple lines of defense possessed by D. occidentalis and the relative value of each line of defense are discussed. It is postulated that aposematic coloration, audible stridulation, and a volatile defensive exudate all function primarily as part of an early warning system enabling a predator to recognize this wasp - with its very algogenic venom - as unpalatable and potentially dangerous.

Dasymutiila occidentalis (L.) is a large, conspicuous mutillid wasp potentially susceptible to the predatory advances of a variety of vertebrates and invertebrates. The apterous diurnally active female wasps, parasites of larger Hymenoptera, engage in extensive searching activities near the ground surface for potential hosts. Such long daily exposures over the lifespan of at least several months add up to an extended period of time in which these wasps are susceptible to predation. In spite of the female's inability to fly and its low reproductive potential, this mutillid is a fairly common species, suggesting that it possesses an effective array of defensive adaptation~ and behaviors. The paucity of more than anecdotal reportings of the defensive mechanisms of mutillid wasps prompted us to look in depth at the various defensive capabilities of this species. The results of this investigation are reported in the present paper.

METHODS AND MATERIALS

Females and males of D. occidentalis were captured during July and August of 1974 and 1975 in Clarke Co., Georgia. Female wasps were maintained in the laboratory in plastic tubs containing sand and were provided with water and honey. Male wasps were utilized only for the preparation of extracts for chemical investigations; in all other studies only female wasps were used.

100 JUSTIN O. SCHMIDT AND MURRAY S. BLUM

Chemical analysis Methylene chloride extracts of both heads and thorax-abdomens of 74 male D.

occidentalis were prepared and stored at -20 ~ until analyzed. Similarly prepared extracts from female wasps were analyzed immediately.

Gas-liquid chromatographic analyses were performed isothermally on a Tracor MT - 220 gas chromatograph equipped with flame ionization detectors and 183 • 0.65 cm glass columns. Columns used were: 3% OV-17 on Chromosorb G AW- DMCS 100/120 mesh (65~ 10% SP-1000 on Supelcoport 80/100 mesh (80~ 10% Carbowax 20 M on Chromosorb W AW-DMCS 100/120 mesh (80~ and 5% SE-30 on Chromosorb W AW-DMCS 60/80 mesh (65~ Combined GLC-mass spectro- scopic analyses were performed using an LKB 9000 instrument with a 183-cm column of 1% OV-17 on Supelcoport 80/100 mesh, programmed from 50 ~ to 300 ~ at a rate of 10~

Quantitative analyses of the volatile components were determined by cutting out the area under each chromatographic peak and weighing it. Ten replicate injections on the OV-17 column maintained at 65 ~ were analyzed.

Sodium borohydride reduction was carried out by adding 20 I~1 of extract to 200 lai of ethanol containing 25 mg of NaBH4. The solution was centrifuged and the supernatant injected onto the OV-17 column at 65~ Standard compounds included 4-methyl-3-heptanone (Aldrich Chemical Co.) and 4-methyl-3-heptanol (Chem. Samples Co.).

Acoustical recordings Recordings of female D. occidentalis were made in the soundproof chamber

maintained by the USDA laboratory in Gainesville, Florida. Recordings of wasp stridulations were made both when a wasp was held with forceps and when it was free but harassed. Amplitude versus time osciliograms and frequency profiles were prepared.

Analysis of cuticule strength Various insects which are considered hard shelled and/or are related taxonomic-

ally to mutiUids, were tested for resistance to a crushing force. A Hanson Cook-O- Meter (Model No. 1310) was adapted to measure the force necessary to crush various body parts of the dried insect specimens. Analysis consisted of the measurement of the force necessary to induce skeletal collapse. Each analysis was replicated ten times.

Cursorial speed Running speeds of D. occidentalis females were determined at a temperature of

25 ~ on ground lightly covered with pine needles. Wasps were repeatedly released and "chased" with fingers, during which time distances and stopwatch times were recorded. A total of ten uninterrupted runs was analyzed.

Predator responses The following potential predators of D. occidentalis were maintained under

ADAPTATIONS AND RESPONSES OF DASYMUTILLA TO PREDATORS 101

laboratory conditions: the red imported fire ant (Solenopsis invicta Buren), the Western harvester ant (Pogonomyrmex owyheei Cole), the Chinese mantid (Tenodera aridifolia sinensis Saussure), wolf spiders (Lycosa spp., Geolycosa sp., L ycosa carolinensis Walck.), tarantulas (Dug esiella hentzi (Girard) and Aphonopelma sp.), the Cuban anole (Anolis sagrei sagrei Duneril & Dibron), the Carolina anole (Anolis carolinensis Voight), the six-lined race runner (Cnemidophorus sexlineatus (L.)), the Florida scrub lizard (Scelophorus woodi Stejneger), the common starling (Sturnus vulgaris L.), and the gerbil (Meriones shawi Duvernoy). Predation responses were measured by observing the responses of both the wasp and the predator when the former was introduced into the environment of the latter. After each encounter with a wasp, the predator was given a German cockroach (Blattella germanica (L.)) or another previously demonstrated palatable insect in order to determine if the predator was hungry. With the possible exception of the lizards, which were field captured, all predators were naive. Results of all experimental encounters, in which the subsequently introduced palatable prey were not eaten, were discarded. In the case of fire ants, laboratory experiments were supple- mented by field tests.

The deterrent value of the major component of mutillid mandibular gland secretion was determined by placing filter paper flags or palatable prey treated with I lal of the compound in the immediate vicinity of potential predators and recording their subsequent reactions.

R E S U L T S

Chemical analysis Preliminary gas chromatography (GC) of cephalic extracts of D. occidentalis on

various columns including SE-30, OV-17, Carbowax 20 M, and SP-1000 demonstrated the presence of one major volatile component, two components in moderate amounts and several trace components. GC-MS analysis of the major component yielded the following prominent peaks: m/e (relative intensity) mol. ion 128 (3), 99 (8), 86 (34), 71 (61), 57 (100), 43 (92), and 29 (52). This spectrum was identical to that of 4-methyl-3-heptanone (McGurk et al., 1966). The identity was further confirmed by simultaneous co-injection of 4-methyl-3-heptanone and the wasp extract. Peak 1 remained sharp and symmetrical while additively increasing in size.

Head extracts of males gave GC patterns identical to those given by female extracts, but contained a much smaller amount of material per wasp. Dissections of the individual sexes revealed the presence of well-developed mandibular gland reservoirs which contained yellowish fluid. Crushing a gland yielded the same characteristic sweet ketonic odor as was detected when a wasp was restrained. GC analysis showed extracts of the mandibular gland to be identical with whole head extracts. GC analysis of the extract of the rest of the body revealed no low-boiling constituents.

Reaction of D. occidentalis extract with sodium borohydride resulted in a loss of the peak corresponding to 4-methyl-3-heptanone and the appearance of a peak


with the same retention time as 4-methyl-3-heptanol. The largest minor constituent did not react with the reagent while the smaller component disappeared. The chemistry of the unidentified compounds is currently under investigation.

Quantitative analysis of the three reproduceable peaks of D. occidentalis gave the following approximate values for percent of total low-boiling volatiles: peak 1 (4-methyl-3-heptanone) 72%; peak 2, 20%; peak 3, 8%.

Sound production When disturbed, both males and females of D. occidentalis produce a sound

audible to the human ear at a distance greater than a meter. Although stridulitra, series of transverse striations, and ridge-like plectra are present on the second through sixth metasomal (= apparent abdominal) tergites (Hermann & Mullen, 1974), our observations indicate the sound is produced mainly by the movement of the stridulitrum on the third metasomal tergite across the plectrum on the posterior margin of the second tergite. Because acoustical recordings of a female wasp held with forceps yielded frequency and amplitude profiles essentially identical to those obtained with a wasp allowed to run free while being harassed, only results obtained with the latter are presented.

Fig. 1 represents an amplitude versus time oscillogram recorded at a slow speed. The variable nature of the signal amplitudes, the irregular time durations of the pulse trains (the closely spaced series of pulses which combine to form each individual "chirp"), and the uneven temporal spacings of the individual pulse trains, is clearly seen. Two single pulse trains are expanded in Fig. 2. Again variability in amplitudes and temporal spacings are evident. Furthermore, the number of individual pulses per pulse train derived from sixteen consecutive trains, ranged from 31 to 44 with an average of 36.1 + 3.27. Fig. 3 represents four individual pulses taken from different parts of pulse trains and recorded on a greatly expanded time scale. Fig. 4 represents a frequency profile from 0 to 32 kilohertz (kHz) for several pulse trains. Fig. 4A represents one direction of the metasomal movement necessary to produce sound and Fig. 4B represents the other direction. As can be seen, no narrow frequency ranges predominate: the intensity gradually increases to a maximum from 4 to 8 kHz and trails off from there to 45 kHz, the sensitivity limit of the apparatus. Background noise was uniform across the profile and was considerably lower than signal level.

Cuticular resistance to crushing The results of forces applied across the mesothorax of various dried insect

specimens, are recorded in Table I. The forces required to crush the head and abdomen were less than those required to crush the thorax. For D. occidentalis, for example, the crushing force applied antero-posteriorly by the bar, placed horizontally across the frons, was 19.5 + 8.3 Newtons and the dorso-ventrally applied force to crush the second metasomal segment, was 14.5 + 3.0 Newtons.

During the crushing tests, it became evident that the actual strength of the cuticle was only part of the protective value of the integument. Excluding appendages, all external integumentry parts of D. occidentalis were extremely

A D A P T A T I O N S A N D R E S P O N S E S O F D A S Y M U T I L L A T O P R E D A T O R S 103

rounded and slippery. These factors frequently resulted in the specimens of this species slipping from within the apparatus during the testing process. No other species exhibited this "slipping" tendency.

Running speed From ten trials using female D. occidentalis at 25 ~ the mean ground speed was

13.8 + 3.0 cm/sec, or roughly 0.5 km/hr.

i i �9 i . L I II I i , t . , I ~ t,m b * I n d L d d l , b l ~ , A I d l J , & a l h d i . J n l ~ i ~ a _ _ I I I d lL I Id i l l i l l I d l t l l l l l l n i i l i i B l i i l _ _ N i n i _

.l See Fig. 1. Oscillogram of stridulatory series of pulse trains produced by D. occidentalis.

i i i l n i l i i i i i i l l l l l l l l l l l l l l u m m n l l l l U U l l n m m l l l i i l m I n

N u

m

i m , l m u ~ , m - ~ q

H n i l ~ ~ 1 n m l

Fig. 2. (A) Oscillogram representing the expansion of the pulse train produced by movement of the metasoma (=apparent abdomen) in one direction; (B) oscillogram of the pulse train produced by

movement of the metasoma in the opposite direction.

1 0 4 JUSTIN O. S C H M I D T AND MURRAY S. BLUM

Predator responses The results of encounters of D. occidentalis and hungry predators are shown in

Table II. Because many of the observational results obtained from these predator studies are difficult to quantify, brief descriptions of the encounters are presented.

When attacked by small numbers of ants, D. occidentalis removes the ants by

I 1 m sec

I

Fig. 3 (A--D). Oscillograms representing expansions of four separate pulses.

I I I

2i

4 8 12 16 20 24 28 F R E Q U E N C Y ( K H Z )

Fig. 4. Frequency profiles of D. occidentalis stridulation. (A) The average of several pulse trains produced by movement of the metasoma in one direction; (B) the average of several pulse trains

produced by movement of the metasoma in the opposite direction.


Insect

TABLE 1

Force in Newtons necessary to crush the thorax of selected dried insects

Common name Approx. Force to dried weight crush thorax2, 3

in g'

Force per gm dry weight to crush thorax 2

Dasvmutilla occidentalis (females) velvet ant 0.112 27.8 _+5.3 Scolia dubia (Say) scoliid wasp 0.083 11.5 _+ 3.2

Apis mellifera L. (workers) honey bee 0.020 2.44 _+ 0.36 Vespula maculata (L.) (queens and workers) bald-faced hornet 0.137 5.74 _+ 1.3 PachylobiuspicivorusGerm.pitcheatingweevil 0.041 11.6 _+3.04

Pseudlucanus capreolus (L.) stag beetle 0.550 14.6 _+ 2.85 26.6 _+ 5.1 Phyllophaga spp. brown June beetles 0.169 5.25 _+ 1.1 31.1 _+ 6.5

5.30_+0.735 31.4_+ 4.3

6.83 -+ 2.14 40.4_+ 12.0

7.50+0.52 ~,5 44.3-+ 3.1

' Average weight o f ten dried specimens Mean force _+ one unit standard deviation

3 Force directed laterally across mesothorax unless otherwise specified 4 Force applied to prothorax 5 Force applied dorso-ventrally

247.0-+ 47.0 139.0_+38.0 123.0_+ 18.0

32.8+ 9.7 284.0 + 73.0

means of rapid scraping movements of the legs. During these attacks the wasp continues walking over the surface in its usual manner. However, when attacked en m a s s e by ants, the wasp simultaneously increases locomotor rate and ant removal activities. In natural surroundings, these two responses are sufficient to allow the wasp to escape from attack by large numbers of formicids, but in the confinement of laboratory containers which housed ant colonies, the wasp was rapidly overwhelmed. At this point it attempted to stridulate, but made no attempt to sting or to bite. After being immobilized for 5 min by a great mass of clinging ants, the observer intervened and removed the wasp (along with up to 39 attached ants). Such wasps rapidly freed themselves of clinging ants and returned to their quotidian behavioral patterns. In no case was a wasp observed to be either temporarily or permauently injured as a result of these encounters. In fact, all wasps were alive and healthy 6 weeks after encounters, a remarkable feat in the light of the extreme pugnacity of the red imported fire ant, one of the most highly successful predators which we have examined.

The behavioral response of D. occ identa l i s to spider attacks depended upon the severity of the attack. When the spider's attack was limited to no more than surface contact, the wasp responded by ceasing locomotor activity, drawing the legs and antennae inwards, and stridulating very loudly. When the spider actually grabbed the wasp and attempted to crush it, the wasp struggled to escape, attempted to sting and bite, and stridulated when able. No small spider (under 25 mm length) ever killed or injured a wasp, even though spiders of this size routinely


Predator

TABLE 1I

Results of encounters between hungry predators and D. occidentalis

Trials/ Number of Predatory Wasp injury 2 and comments predator t predators attack?

lnsecta: S. invicta (red imported fire ant) S. invicta

P. owyheei (Western harvester ant) 7". a. sinensis (Chinese mantid)

Arachnida: Lycosa spp (8, 13, 15, 3--6 5 yes/no 17, 17 mm) (wolf spider) L. carolinensis (25 mm) 10 1 yes (Carolina wolf spider) Geolycosa sp. 4 1 yes/no (burrowing wolf spider) D. hentzi (16--20 mm) 2 10 yes/no (tarantula spiderlings) D. hentzi & Aphonopelma 1 3 yes sp. (25, 40, 60 mm) (American tarantulas)

Reptilia: A. s. sagrei 2 6 yes (Cuban anole) A. c. carolinensis 2 2 no (Carolina anole) S. woodi 3 4 yes/no (Florida scrub lizard)

C. sexlineatus 2 1 no (race runner lizard)

Ayes: S. vulgaris 2 1 no (common starling)

Mammalia: M. shawl 2 8 yes/no (gerbil)

2 2 colonies yes 0 wasps placed in lab colonies of ants

6 2 colonies yes 0 wasps placed on top of mounds in the field

2 2 colonies yes 0 wasps placed in lab colonies of ants

4 1 yes 0

0 Ist attack most aggressive, often not attacked thereafter

0 aggressiveness of attack decreased with time

0 only initial two encounters resulted in an attack

0 1st attack most aggressive, often not attacked thereafter

0,2 one wasp eaten

0 each lizard attacked only on the initial encounter

0

0 largest lizard attacked twice, one lizard once, and two never attacked

0

0,2 four gerbils frightened by wasp, two attacked once, two attacked twice; one wasp eaten

Each trial involved the encounter of one predator with a wasp on a given day. All repeated trials were on different days.

2 Wasp injury scale: O~no injury, 1 =injured but not consumed, 2 =killed and eaten.

A D A P T A T I O N S A N D R E S P O N S E S OF D A S Y M U T I L L A T O P R E D A T O R S 107

consumed yellow jackets (Vespula maculifrons (Buysson)) and harvester ants (Pogonomyrmex badius Latr.). However, the outcome of encounters with larger spiders depended on the aggressiveness of the spider and its mode of attack. Attacks were successful only when the spider forced a chelicera between the wasp's metasomal sclerites and thereby injected its venom. Unless this occurred, attacks were unsuccessful because the chelicerae were unable to pierce any area of sclerotized cuticle. Frequently the spiders exhibited an intrinsic avoidance of a stridulating wasp, with many attacks being abruptly halted when the wasp commenced stridulation. There appeared to be a slow learning process occurring in the case of some spiders as indicated by a progressive reduction in the aggressiveness of their attacks on subsequent trials.

When attacked by a large Chinese mantid, Tenodera aridifolia sinensis, a mutillid wasp was so completely immobilized that, often, even its ability to stridulate was repressed and movement was limited to the extrusion and retraction of its sting. During this time the mantid attempted, unsuccessfully, to chew through various parts of the thorax and abdomen. These encounters, which lasted from 10--60 sec, were terminated abruptly when the mantid threw the wasp aside. In at least one encounter the mantid was observed to be stung in the mouth; possibly, this might have also occurred in encounters involving other wasps. The mantid routinely consumed workers of Vespula and Pogonomyrmex, including sting and poison apparatus.

When placed in the cages of any of the lizard species, the mutillids were generally unaware of the lizards until attacked. When grasped by a lizard, a wasp immediately stridulated and attempted to sting. However, because the lizards frequently released the wasps upon contact and exhibited no apparent signs of distress, the sting was apparently not always essential in causing the release. The cuticle strength, the stridulation and/or the chemical secretion may have contributed to causing the release. Once released, the wasp stridulated while fleeing. After the initial attack, a lizard usually did not attack again, and the wasp, which continued to move at an accelerated rate, stridulated when movement occurred nearby.

When placed in a cage with a starling, Sternus vulgaris, the wasp crawled around as would be expected in a new environment. The starling, however, noted the wasp but made no effort to move towards or attack it. After trials, the bird aggressively attacked mealworms.

Gerbils (Meriones shawi), highly insectivorous mammals which consumed, among other insects, pentatomid bugs and frequently Campanotus ants, exhibited variable behavior toward D. occidentalis. Often on the initial encounter, the gerbil attacked the wasp and quickly released it. Stung gerbils withdrew from the wasp and vigorously scratched their mouths. Afterwards, a close approach to the stridulating wasp usually resulted in the gerbil fleeing the area. Despite the powerful sting, the wasp's defenses were not always adequate. After first being stung, one gerbil returned to attack again, this time consuming the wasp without receiving another sting.


Responses of predators to 4-methyl-3-heptanone Predators displayed various responses to filter paper flags impregnated with 1 ~tl

of 4-methyl-3-heptanone. When a flag was placed on top of a drop of honey upon which fire ants were feeding, the ants moved about 1 cm away and waited for the ketone to dissipate before returning to the food. On the other hand, P. badius workers exhibited typical alarm behavior in the presence of the ketone which is their natural alarm pheromone (McGurk et al., 1966). Neither the spiders nor the mantid exhibited any obvious responses to the nearby ketone-treated flags. Of the vertebrates, lizards generally showed weak aversion, the gerbils appeared to show little interest and the starling was not tested. When 1 lal of 4-methyl-3-heptanone was placed on a palatable prey, predator responses varied considerably. All the arthropod predators ate the prey without exhibiting great aversion to the chemical. A. carolinensis, though eating the prey, exhibited atypical signs of discomfort. S. woodi demonstrated a clear aversion, often rejecting prey freshly covered with 0.3--0.5 ~tl of 4-methyl-3-heptanone. The gerbils typically grabbed the ketone-covered prey, plowed their noses through the sand in the cage, and returned to eventually consume the prey. A. sagrei, C. sexlineatus, and the starling were not tested.

DISCUSSION

D. occidentalis appears to be protected from predation by at least three direct adaptations: a mobile sting capable of extension more than ! cm (Hermann, 1968) associated with the secretion of an algogenic venom; an extremely strong, rounded, slippery cuticle capable of withstanding a much greater crushing force than the hard beetles and other stinging Hymenoptera tested; and the ability to effect a rapid cursorial escape from all but the largest of predators.

To complement the above defensive systems, the insect possesses an aposematic color pattern, a form of early warning which may indicate to the predator that this wasp is unpalatable, and can audibly stridulate and/or release an exocrine secretion.

4-Methyl-3-heptanone is the first exocrine product reported from a mutillid wasp species and, excluding venoms, is the first reported defensive compound from the non-social aculeate wasps. This ketone is widely distributed among ants, including species from at least five subfamilies (McGurk et al., 1966; Moser et al., 1968; Blum et al., 1968; Crewe & Blum, 1972; Fales et al., 1972; Duffield & Blum, 1973; Blum & Brand, 1972; Blum et al., 1973) and is also produced by opilionids (Bium & Edgar, 1971; Meinwald et al., 1971), an arthropod group phylogeneticaily distant from either ants or mutillid wasps.

4-Methyl-3-heptanone apparently functions as a primary constituent of a defensive secretion. This role is supported by our results with lizards as well as by literature precedents in which the ketone has been shown to be an integral part of defensive secretions of opilionids (Blum & Edgar, 1971; Meinwald et al., 1971), and ant alarm pheromones. Alarm pheromones in ants are felt to be derivations of ancestral defensive secretions (Blum, 1971). Since nearly 30% of the secretion of


D. occidentalis consists of unidentified constituents, the deterrent potency of this exudate against selected predators may be considerably greater than that possible with 4-methyl-3-heptanone alone.

Both sexes of D. occidentalis stridulate readily when threatened. The sound produced consists of a broad spectrum of frequencies produced in a non-uniform pattern. Such a sound structure indicates that the stridulation does not serve as an intraspecific communication system for this solitary insect. The literature abounds with examples in which broad frequency, non-uniform sounds have been suggested as components of defensive repertoires (cf Haskeli, 1957; Alexander et al., 1963; Ciaridge, 1974; Eisner et al., 1974; Freitag & Lee, 1972; Blest, 1964; Frauca, 1969; Crowcroft, 1957; Roth & Hartman, 1967; Alexander, 1958; Wheeler et al., 1970; Dunning & Roeder, 1965; Lane & Rothschild, 1965). In most of the above citations of defense by sound, the candidate arthropods also possess other deterrent systems such as irritating secretions, offensive tastes, or painful stings. D. occidentalis, an insect possessing an extremely algogenic venom, is perhaps a classic example of a highly protected insect gaining further protection through the vehicle of sound. The sound produced by D. occidentalis as well as other arthropods possessing deterrent attributes very likely functions as an aposematic acoustic warning: a signal to potential predators that the insect producing the sound is unpalatable.

One might well expect ants, predaceous beetles, and perhaps spiders to be major invertebrate predators of mutillids. The extremely hard cuticle, the rapid cursorial abilities, stridulation and perhaps, as a last resort, the sting would appear to constitute the primary defenses against these arthropod predators. Ants, in general, are effectively deterred by the mutillid's hard slippery cuticle and its running speed; some spiders appear to be deterred by the stridulation (perhaps no other lines of defense are needed against these arthropods), others may be deterred mainly by the cuticular armanent and slipperiness combined with swift escape. The praying mantis, probably an unnatural predator in the mutillid environment, could not chew through the hard, slippery cuticle and may eventually be repelled by the sting.

Vertebrates pose quite a different threat to mutillid survival. The sting plus the cuticular strength and slipperiness appear to provide the basic protection against these animals. The cursorial abilities are of value mainly insofar as they may allow the wasp to escape either by entering a hole in the ground or by running to other cover. Because most vertebrate predators possess exceptionally well-developed senses of vision, hearing, and smell (especially mammals) as well as highly developed central nervous systems, they are capable of perceiving and associating warning colors, sounds and odors with potentially painful or harmful consequen- ces. D. occidentalis, whose sting is very painful to vertebrates, is therefore ideally suited to gain further protection from these predators by the use of its aposematic color, sounds, and odors.

From the above discussion, the value to D. occidentalis of possessing numerous defensive adaptations, is apparent. The wasp may at times appear "over

1 10 JUSTIN O. SCHMIDT AND MURRAY S. BLUM

protected"; however, a diversity of predators requires a diversity of protective mechanisms. Furthermore, the effect of one defensive system is usually synergized by the action of another. Indeed, a diversity of defensive strategies may also insure better wasp survival through the conservative utilization of various defensive lines only as mandated by the circumstances.

We thank Thomas Walker for assistance with the acoustics, Joe Benner, ARS- USDA Gainesville, Fla, for making the acoustical recordings, Barbra Green for lending us the strength-testing equipment, and especially Henry Fales for criticisms and suggestions of chemistry. We are grateful to Robert Matthews, Sean Duffey, John MacDonald, and Ching Tsao for critical reviews of this manuscript.

ZUSAMMENFASSUNG

ANPASSUNGEN UND REA KTIONEN VON DASYM UTI LLA OCCI DENTA LIS (HYMENOPTERA : MUTILLIDAE) GEGENUBER R,4"UBERN

Die Mutillide Wespe Dasymutilla occidentalis besitzt verschiedene Anpassungen und zeigt eine An- zahl von Reaktionen, die ffir die Verteidigung von Wert sind: ein langer, beweglicher Stachel mit starkem Gift, eine starke, runde und glatte Kutikula, die M6glichkeit, sehr schnell und ausweichend zu laufen sowie ein aposematisches Warn-Farben-Muster. Desweiteren ist sie ffihig, einem Angriff mit einem knisternden Ton entgegenzutreten sowie ein chemisches Sekret abzugeben oder auch beides, lm Labor durchgeffihrte Zusammenst6ge zwischen vertebraten und invertebraten Raubern und D. occidentalis beweisen klar den Wert der ganzen Anzahl von Verteidigungsmechanismen f/Jr die Wespe.

Diese Zusammenst6ge liefern auch einen Einblick in die raison d'etre der vielseitigen Verteidigung. Die st~rkere Kutikula und ihre Glfitte, Haupteigenschaften zum Uberleben gegen R~iuber, funktionie- ren gleichzeitig mit dem Stich zum Schutz gegen Vertebraten und mit schneller Fluchtm6glichkeit zum Schutz gegen die meisten lnvertebraten. Das akustische GerS.usch scheint eine Hilfsverteidigung zu sein wenigstens gegen einige Spinnen und vermutlich auch gegen einige Vertebraten. Die chemische Sekretion, die haupts/ichlich aus 4-methyl-3-heptanon besteht, scheint m6glicherweise direkten Ver- teidigungswert gegen einige Eidechsen zu haben und funktioniert h6chstwahrscheinlich haupts/ichlich im Zusammenhang mit der roten und schwarzen aposematischen F/irbung und dem knisternden Ton als vielseitiges Warnsystem, das ffihig ist, durch Signalisieren allen potentiellen vertebraten R~ubern mit- zuteilen, dag dieses Insekt ungeniegbar ist.

REFERENCES

ALEXANDER, A. J. (1958). On stridulation of scorpions. Behaviour 12 : 339--352. ALEXANDER, R. D., MOORE, T. E. & WOODRUFF, R. E. (1963). The evolutionary differentiation of stri-

dulatory signals in beetles. Anim. Behav. 11 : 111--115. BLEST, A. D. (1964). Protective display and sound production in some new world arctiid and ctenuchiid

moths. Zoologica 49 : 161--181. BLUM, M. S. (1971). Dimensions of chemical sociality. Proc. 2nd Int. IUPAC Congr. in A. S. TAHORI

(ed.): Chemical Releasers in Insects Vol. 3, Tel-Aviv, Israel. p. 147--162. BLUM, M. S. & BRAND, J. M. (1972). Social insect pheromones: their chemistry and function. Am. Zool.

12 : 553--576. BLUM, M. S. & EDGAR, A. R. (1971). 4-Methyl-3-heptanone: identification and role in opilionid exo-

crine secretions. Insect Biochem. 1 : 181--188. BLUM, M. S., PADOVANI, F. & AMANTE, E. (1968). Alkanones and terpenes in the mandibular glands of

Atta species (Hymenoptera: Formicidae). Comp. Biochem. Physiol. 26 : 291--299.

ADAPTATIONS AND RESPONSES OF DASYMUTILLA TO PREDATORS 11 I

BLUM, M. S., WHEELER, J. W., DUFFIELD, R. M. & FRANKIE, G. W. (1973). Unpublished. CfBlum, M. S. (1973). Comparative exocrinology of the Formicidae. Proc. VII Congr. IUSSI. London pp. 23--40.

CLARIDGE, M. F. (1974). Stridulation and defensive behaviour in the ground beetle, Cychrus caraboides (L.). J. Ent. Set. A gen. Ent. 49 : 7--15.

CrtEWE, R. M. & BLUM, M. S. (1972). Alarm pheromones of the Attini: their phylogenetic significance. J. Insect Physiol. 18 : 31--42.

CROWCROFT, P. (1957). The Life of the Shrew. Max Reinhardt. London. 166 p. DUFEIELD, R. M. & BLUM, M. S. (1973). 4-Methyl-3-heptanone: identification and function in Neopone-

ra villosa. Ann. ent. Soc. Am. 66 : 1357. DUNNING, D. C. & ROEDER, K. D. (1965). Moth sounds and the insect catching behavior of bats.

Science 147 : 173--174. EISNER, T., ANESHANSLEY, D., EISNER, M., RUTOWSK1, R., CHONG, B. & MEINWALD, J. (1974). Chemi-

cal defense and sound production in Australian tenebrionid beetles (Adelium spp.). Psyche 81 : 189--208.

FALES, H. M., BLt;M, M. S., CREWE, R. M. & BRAND, J. M. (1972). Alarm pheromones in the genus Manica derived from the mandibular gland. J. Insect Physiol. 18 : 1077--1088.

FRAt;CA, H. (1969). Australia's giant barking spider. Entomologist 102 : 259--260. FREITAG, R. & LEE, S. K. (1972). Sound-producing structures in adult Cicindela tranquebarica (Coleo-

ptera: Cicindelidae) including a list of tiger beetles and ground beetles with flight wing files. Can. Ent. 104 : 851--857.

HASKELL, P. T. (1957). Stridulation and its analysis in certain Geocorisae (Hemiptera: Heteroptera). Proc. zool. Soc. London 129: 351--358.

HERMANN, JR. H. R. (1968). The hymenopterous poison apparatus. IV. Dasymutilla occidentalis (Hyme- noptera: Mutillidae). J. Georgia ent. Soc. 3 : 1--10.

HERMANN, JR. H. R. & MULLEN, M. A. (1974). Stridulatory structures in Dasymutilla occidentalis (Hymenoptera: Mutillidae). J. Georgia ent. Soc. 9 : 204--205.

LANE, C. & ROTHSCHILD, M. (1965). A case of Mullerian mimicry of sound. Proc. R. ent. Soc. Lond. 40 : 156--158.

McGuRK, D, J., FROST, J., EISENBRAUN, E. J., VICK, K., DREW, W. A. & YOUNG, J. (1966). Volatile compounds in ants: identification of 4-methyl-3-heptanone from Pogonomyrmex ants. J. Insect Physiol. 12 : 1435--1441.

MEmWALD, J., KLUGE, A. F., CARREL, J. E. & EISNER, T. (1971). Acyclic ketones in the defensive secretion of a "daddy longlegs" (Leiobunum vittatum). Proc. natn. Acad. Sci. U.S.A. 68 : 1467--1468.

MOSER, J. C., BROWNLEE, R. C. & SILVERSTEIN, R. (1968). Alarm pheromones of the ant Atta texana. J. Insect Physiol. 14 : 529--535.

ROTH, L. M. & HARTMAN, H. B. (1967). Sound production and its evolutionary significance in the Blat- taria. Ann. ent. Soc. Am. 60 : 740--752.

WHEELER, J. W., CHUNG, R. H., OIL S. K., BENFIELD, E. F. & NEFE, S. E. (1970). Defensive secretions of cychrine beetles (Coleoptera: Carabidae). Ann. ent. Soc. Am. 63 : 469--471.

Received for publication: October 11, 1976