628 BEHAVIOUR, 30, 2 198A

(Lighthill 1952). Because of the effects of the half-muting Lighthill, M. J. 1952. Oq sound generated aerodynami- operation on the quality of courtship song, we surmise cally. I. General theory. Proc. Roy. Soc. A., 211, that G. chopardi makes use of both of these variables to -564-587. - . - - . . produce both hisses and whistles with the same pair of Masters, W. M. 1980. Insect disturbance stridulation: spiracles. The courtship whistles of half-muted males are characterization of airborne and vibrational com- consistently higher in frequency (by p to 1 kHz) than ponents of the sound. J. comp. Physiol., 135, the same males' pre-operative whisJes; they are also 259-263. interrupted frequently by brief lapses into broad-band Nelson, M. C. 1979. Sound production in the cockroach, noise. If males have a fixed motor programme for the Gromphadorhina portentosa : the sound-producing abdominal compression that forces air through the apparatus. J. comp. Physiol., 132, 27-38. tracheal horn, the rate of air flow should be higher than Nelson, M. C. & Fraser, J. 1980. Sound production in the normal through the single spiracle that is capable of cockroach, Gromphadorhina portentosa: evidence opening, causing a shift toward a higher frequency in the for communication by hissing. Behav. Ecol. resulting sound. If whistles are produced with the valve Sociobiol., 6, 305-314. partly closed and hisses are produced with the valve fully Pipher, R. D. & Morris, G. K. 1974. Frequency modu- open, the higher rate of flow in half muted males might lation in Conocephalus nigropleurum, the Black- well lead to an occasional 'blow-out' of the valve, ~idedmeadow katydid (Orthoptera: Tettigoniidae). resulting in abrupt lapses from whistle into hiss (as in Can. Entom., 106,997-1001. Fig. 1, Plate I: 'R.Sp. Blocked' row 2). Young, D. 1972. Neuromuscular mechanisms of sound Why does G. chopardi u$e both whistles and hisses? We production in Australian cicadas. J. comp. suggest two explanations. First, the disturbance hiss, like Physiol., 79, 343-362. many defensive sounds, has a broad frequency spectrum which probably is advantageous to the emitter, since a (Received 19 August 1981 ; revised 17 December 1981 ; broad range of predators can detect the sound (Masters MS. number: AS-153) 1980). Second, preliminary studies show sound produc- tion is used in very similar ways in the courtship of portentosa and chopardi. The great difference in the Role of the Elephant Pinna in Sound Localization frequency characteristics of their courtship sounds may Because of its unique size and mobility, the elephant thus represent a species isolation mechanism (both pinna has been the subject of studies aimed at determining species have been collected from the same regions of its functional significance (for a review, see Buss & Estes Madagascar although direct evidence for sympatry is 1971). The results of these studies have suggested two lacking; Chopard 1950). This unusual divergence in the possible roles for the elephant pinna. First, large differ- sound signalling systems of two such closely related ences in arterio-venous .temperature in the pinna, coup- svecies undoubtedlv owes its emergence to the nature of led with the observation that elephants increase the rate ~rom~hadorhina'smethod of souna production, in which at which they flap their pinnae as ambient temperature the potential for frequency modulation is inherent in the increases, suggest that the pinna plays a role in regula- design. ting body temperature. Second, the observation that We wish to thank the students of the MBL (Woods elephants extend their pinnae rigidly from their heads Hole) Neural Systems & Behavior Course, class of 1979 in alarm or threat situations has been interpreted to (C. Derby and T. Flanagan especially), without whom mean that they use the pinnae to alert nearby elephants this project might never have begun. R. Hoy, R. Murphey, or to warn intruders. Thus, the elephant pinna has been and J. Paton made many helpful comments on earlier implicated in both thermoregulation and communication. drafts of this MS. This research was supported by Recently, we have had the opportunity to test the NINCDS Grant NS16323 to MCN. hearing capacities of a 7-year-old wild-born female JEANFRASER* Indian elephant (Elephas maximus) at the Ralph MARGARETC. NELSON? Mitchell Zoo in Independence, Kansas. The tests given the animal included the determination of absolute, *Department of Biology, sound-localization, and frequency-discrimination thres- Boston University, holds. During these tests it was noted that the Boston, MA 02215. elephant extended its pinnae when attempting to localize brief sounds, but did not extend them during the non- ?Section of Neurobiology & Behavior, localization tests. The purpose of this paper, then, is to Cornell University, describe these observations and to suggest that the ele- Zthaca, NY 14853. phant's pinna may play an important role in the local- ization of sound. - References The details of the behavioural procedure used to test I Chopard, L. 1950. Sur I'anatomie et le dbveloppement the elephant have been described elsewhere (Heffner & d'une blatte vivipare. Znt. Congr. Ent., 8 Heffner 1980. 1981). Brieflv. the elevhant was brought (Stockholm 1948), 218:222. into a room (6.7 x 5.6 x 4:6 m) and tethered in front Dumortier, B. 1963. The phys~calcharacteristics of sound of a response panel which contained three buttons emissions in arthropods. In: Acoustic Behaviour of mounted in a horizontal row. The animal was trained to (Ed. by R. -G. Busnel), pp. 346-373. make an 'observing' response by pressing the centre London : Elsevier. button with its trunk, a response which initiated a trial Elsner, N. 1974. of sound production in and was followed by the presentation of an auditory cue, Gomphocerine grasshoppers - (Orthoptera : In the sound-localization tests, the animal was then re- ' ' Acrididad I. Sona vatterns and stridulatow quired to press the left or right button if a sound cmca movements. J. corn;. ~h~siol.,88, 67-102. from a loudspeaker located to the left or right of bhe Elsner, N. & Popov, A. 1978. Neuroethology of acoustic animal, respectively. A comt response was rew&dacll communication. Adv. Insect Physiol., 13, 229-355. by dispensing 30 ml of a fruit-flavoured sugar s01t.doim SHORT COMA

into a small trough located below the centre button. In signs of reduced motivation by occasionally ceasing to the absolute-threshold tests, the animal was required to [perform and playfully pushing on the fenq behind press the centre button and then press the left button if which the loudspeakers were mounted. Though it wn- a sound was presented from a loudspeaker and to press tinued to extend its pinnae on most of the trials, it oc- the right button if no sound was presented. Frequency- casionally failed to extend them or extended only one of discrimination texts were conducted in a similar man- them. Thus, while the poorer performance observed in ner in which, after pressing the centre button, the animal this situation may have been due to reduced motivation, was required to press the left button if a train of tone it is of interest to note that failure to extend the pinnae pulses emitted from a loudspeaker were all of the same was accompanied by poorer performance in localizing 1 frequency and to press the right button if the tone pulses sound. alternated between two different frequencies. Test In summary, the elephant was observed to extend its 1 sessions were conducted during the quiet early morning pinnae during sound-localization tests, but not during hours shortly before the animal's morning feeding time. tests of absolute sensitivity or frequency discrimination. Following initial training, the elephant's absolute Pinna extension was observed to occur for a variety of r thresholds for pure tones from 16 Hz to 12 kHz were different sounds and at both large and small angles. ,, determined. During these tests it was noted that the ele- However, when the animal was insufficiently motivated . phant's pinnae were maintained in the relaxed position to work at a steady pace, it did not always extend its back against its head and neck, with only occasional pinnae and its error rate increased. These observations flapping movements such as to remove flies. However, suggest that elephants extend their pinnae incorder to when the sound-localization tests were given, a very dif- localize sound accurately. I ferent pattern of pinnae movements emerged. During Though it is widely believed that movabb pinnae, A sound-localization testing, the animal extended its such as the elephant's, play an important role in sound pinnae nearly perpendicular to its head just before it localization, it is not clear just how pinna extension en- pressed the centre button. After the centre button was hances the ability to localize sound. While the function pressed and a brief sound had been presented, the ani- of the human pinna has been studied in detail, little is mal then moved its trunk to press either the left or right known concerning the contribution of large movable button, during which time it usually returned its ears to pinnae to sound localization (for reviews, see Butler the normal relaxed position against its head. Following 1975; Shaw 1974). One possibility is that extension of the localization tests, the elephant was tested on fre- the pinnae may enhance the difference between the fre- quency discrimination and was re-tested on its absolute quency-intensity spectra of the sounds reaching the sensitivity to 4 kHz and 8 kHz. In both of these sub- two ears, thereby providing a more distinct localization sequent tests, the animal maintained its pinnae in the cue. Another is that it may reduce the intensity of echoes relaxed position. Thus, the elephant was observed to reaching the ears which otherwise might interfere with display pinna extension only during the sound-local- accurate localization. However, at the present time, the ization tests. effect of pinna extension on the sounds arriving at the Because the elephant was not continuously observed two ears has not yet been determined. during testing, it is not possible to say if it extended its The discovery that the elephant pinna plays a role in pinnae on each of the more than 1000 sound-localization sound localization does not contradict previous ideas trials it received each session. However, we observed concerning its function. The elephant pinna may very consistent pinna extension on trials involving the follow- well be a mechanism for cooling the blood and extension ing stimuli: single clicks, 100-ms bursts of white noise, of the pinnae may serve to alert nearby animals. Indeed, 100-ms bursts of filtered noise with centre frequencies since an elephant actively engaged in localizing a sound from 125 Hz to 8 kHz, and 1Wms tone pips ranging in source is an aroused animal, it would seem natural for frequency from 125 Hz to 8 kHz. In addition, pinna ex- other animals to be alerted by the accompanying pinna tension was observed for angles of separation ranging extension. However, as it has been demonstrated that from 0 to 60". In short, pinna extension was observed for even a pinna as diminutive as that of humans plays a every stimulus and angle used in the sound-localization role in sound localization, it should wme as no surprise tests. In contrast, it was never observed during the de- that the elephant pinna may play an important role in termination of absolute and frequency-discrimination enabling elephants to accurately localize the source of a thresholds which involved frequencies from 16 Hz to sound. 12 kHz and loudspeaker locations of 10" and 30" to the This research was supported by NSF grant BNS left or right of the animal. 5807391, and NIH grant HD 02528, to the Bureau of The conditions under which pinna extension could be Child Research, University of Kansas. observed were further elucidated during a special late- morning session in which the elephant" pinna move- RICKYEHEPFNER ments were photographed. Normally, the animal was HENRYHEPFNER not tested at this time because it had just received food NED STICHMAN* and fresh water and was usually too satiated to perform Bureau of Child Research, reliably. When the animal was allowed to perform a University of Kansas, sound-localization test with single clicks at an angular Parsons, KS, 67357, U.S.A. separation of 20°, it was noted that the animal did not at first extend its pinnae. Furthermore, while the ele- *Ralph Mitchell Zoo, phant performed at a level of 90% wrrect, it was making Independence, KS, 67301, U.S.A. more errors than it normally did on this task. After 30 trials, the angle of separation was reduced to lo0, fol- References lowing which the elephant made two consecutive errors. Buss, I. 0. & Estes, J. A. 1971. The functional signifi- At this point the elephant began to extend its pinnae on cance of movements and positions of the pinnae each trial and its performance rose to near perfect. of the African elephant, Loxodonta africana. During the remainder of this session, the animal showed J. ., 52, 21-27. 1 Y+ t

ANIMAL BEHAVIOUR, 30, 2 I/ 3 '4 , Butler, R. A. 1975. The influence of the external and is unlike, for example, 'kleptogamy', a word which rnmy middle ear on auditory disbriminations. In: of us do not recognize, and which is thus immediately \ Handbook of Sensory Physiology, Vol. V/2 (Ed. by flagged as jargon. W. D. Keidel & W. D. Neff), pp. 247-260. New The compounding of social arguments with arguments York: Springer-Verlag. of adaptive significance is the crux of my concern about Heffner, H. & Heffner, R. 1981. Functional interaural paradoxical jargon. I think the following examples are distance and high-frequency hearing in the ele- fllustrative. At a recent scientific meeting I was asked, phant. J. acoust. Soc. Amer., 70, 1794-1795. Is it rape when avirgin is forced to intercourse?' Similarly, Heffner, R. & Heffnqr, H. 1980. Hearing in the elephant is it rape when a post-menopausal woman is forced to (Elephas maximus). Science, N. Y., 208, 518-520. intercourse? There are two answers to both questions Shaw, E. A. G. 1974. The external ear. In: Handbook of which illustrate my point about paradoxical jargon. Sensory Physiology, Vol. V/1 (Ed. by W. D. Obviously, both are rape - in all of the traditional, Keidel & W. D. Neff), pp. 445-490. New York: psychological, sociological, and legal ways we think of Springer-Verlag. rape. However, in the sociobiological sense of Thornhill, neither of these would be rape, because in the first example (Received 27 August 1981 ; revised 6 November 1981 ; the virgin could be impregnated and her genetic fitness MS. number: AS-154) thereby increased, and because in the second example the fitness of the male who forces a aost-menoaausal female Sexual Terms in : Emotionally Evocative and* to copulate would not be incre&ed, and in-addition, the Paradoxically, Jargon fitness of the menopausal female would not be decreased. The social danger I see is the possibility of transference 'Rape', 'coy', 'cuckoldry', 'adultery', 'homo~exual'~ of the sociobiological idea to the legal realm so that 'harem' are some of the sexual terms used in sociobiology be (Barash 1976, 1977; Abele & Gilchrist 1977; Thornhill legally, victims of rape will said not to have been 1980; and many others). I object to the use of these terms raped. This is not so unlikely. A similar reversal, in which in studies of non-human sociobiology for three reasons: the victim's motives and behaviour are tried, has led to these words used in a sociobiological context are jargon the further victimization of women in police precincts with possible social repercussions; they are more emotion- and courts when they report rapes (Brownmiller 1975). ally evocative and less semantically adaptable than alter- One way to guard against this social danger is to clearly native, descriptive terms; and their use is often sensational separate legal aspects of rape from sociobiological and conducive of an inappropriate prurience. aspects of forced insemination. I suggest that we use a My objection to such sexual terms is associated with descriptive and operational term like 'forced copulation' semantic and definitional problems. These objectionable (see e.g. Gladstone 1979; Burns et al. 1980) or 'forced words are seldom defined and when they are defined, insemination' in discussions of non-human sociobiology agreement on the definitions is hard to reach. 'Rape', for and in discussions of the adaptive significance of the example, has social (Klemmack & Klemmack 1976), phenomenon in humans. I think 'rape' should be used psychological, legal, ideological (Burgess & Holmstrom only in traditional psychological, sociological, and legal 1974; Brownmiller 1975), and now, sociobiological discussions of the phenomenon among people. (Thornhill 1980) meanings. Precise definitions of these My second objection is simply that the emotions evoked sexual phenomena are obviously imperative. by these words suggest important bias similar to those A subtle definitional difficulty surrounds the meaning pointed out by opponents of sociobiology (e.g. Lewontin of these words in studies of adaptive significance. For 1977). In line with their criticisms, Thornhill's hypothesis instance, take Thornhill's (1980) definition of rape. He can be said to reflect current, male-centred, American says or implies (pages 52, 55. 57) that rape is forced in- attitudes as easilv as it reflects the biolom of semination or fertilization, and limits rape to those scorpionflies. In geheral, the suggestion of bias iseasily occurrences of apparent forced insemination or fertiliza- eliminated by unevocative, adaptable, operational, alter- tion that simultaneously enhance the fitness of males and native terms such as 'forced copulation' for rape decrease the fitness of females. This ddfinition is devoid (Gladstone 1979; Burns et al. 1980), 'one-male social unit' of sociological, psychological and moralistic connota- for harem, 'kleptogamy' for cuckoldry (May & Robertson tions. Rape is no longer an act of sexual assault or forced 1980; Gowaty 198l), 'unisexual' for homosexual. These intromission: gametic union or at least the deposition of terms have other advantages. 'One-male social unit' can sperm is also required. Most important, the future easily become 'two-male social unit' or 'multi-male social fitnesess of the putatively raped female and the putative unit', while the idea of a 'double-harem' is at best rapist must also be evaluated. This is a different concept ambiguous. Cuckoldry is kleptogamy as experienced by from the one used commonly for humans to mean any one participant in a threesome of participants. The act of forced intromission or sexual assault. It is more adulterer is one participant; the cuckold and cuckolder than descriptive and operational: this precise and restric- are the others. 'Kleptogamy' is jargon, as are 'adultefer', tive definition allows the study of a mechanism in sexual 'cuckold', and 'cuckolder', but it is eas~ly-~dent~fied selection which overcomes mechanisms of epigamic jargon which has the advantage of clarifying how the selection. In my opinion Thornhill's (1980) examination sexual strategies of the three participants are related; and of aspects of forced insemination in scorpionflies goes a kleptogamy' has no emotional or moral connotations. long way toward a credible theory of the adaptive value Even though I think 'cuckoldry' is an inferior alter- of forced insemination. However, this special socio- native to 'kleptogamy', the interesting defence of the biological use of 'rape' makes the word paradoxical term 'cuckoldry' made by Power et al. (1981) is to be jargon. Jargon is the specialized or technical language of lauded in that they examined the etymological roots of a profession. 'Rape', as used in the sociobiological sense, the word. Because science is culture-bound, etymologies has a new, specialized and technical meaning which is may give important insight into the broader meanings of quite restrictive, yet 'rape', as a word we all recognize, does words, and reveal potential sources of emotional and not make the new connotation obvious. The term 'rape' societal bias. A false argument in favour of the term