William G. Eberhard and J. Mauricio Garcia-C.
Ritual jousting by horned Parisoschoenus expositus weevils (Coleoptera, Curculionidae, Baridinae).
Psyche 103:55-84, 2000.

Full text (searchable PDF, 4200K)
Durable link: http://psyche.entclub.org/103/103-055.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

RITUAL JOUSTING BY HORNED PARISOSCHOENUS EXPOSI- TVS WEEVILS (COLEOPTERA, CURCULIONIDAE, BARIDINAE) ~mithsonian Tropical Research Institute, and Escuela de Biologia, Universidad de Costa Rica, Ciudad Universitaria, Costa Rica; email: weberhar@cariari.ucr.ac.cr
2~pdo. 1 179-2 100, Guadalupe, San Jose, Costa Rica; email: mgarciac@solracsa.co.cr
Males of the weevil Parisoschoenus expositus use their prothoracic horns as weapons in stylized battles with other males over females that are drilling oviposition holes in palm leaves. The unusual sheath-like structures that penetrate deep into the male prothorax function to receive the horns of opponents during battles. Horn size is dimorphic with respect to body size, and small and large males also differ behav- iorally. Small males that have mated with a drilling female are some- times able to impede a large male's access to the female until after she has oviposited, but they are not able to take over females from larger males.
Beetle horns are extremely diverse in size and shape (e.g., Arrow 195 1; Eberhard 1979). Despite occasional claims to the contrary (Mdler 1992), observations of their use in natural contexts suggest that they function as weapons in battles between conspecifics (Moron 1976; Bechtle 1977; Eberhard 1977, 1979, 198 1, 1987; Palmer 1978; Brown and Siegfried 1983; Otronen 1988; Connor 1988; Siva-Jothy 1989; Rasmussen 1994; Emlen 1994, 1997), rather than as visual display devices as do the horns and antlers of some ungulates (Geist 1966, 1978; Lincoln 1994). Three common functional designs have been documented among beetle horns: a dorso-ventrally mobile head horn which serves as a lever to lift the opponent and (in some species) to Manuscript received 28 October 1998.
55
Paths 103:55-84 (MOO). htlp$psychc mtclub.ore/l03/10->-055.himl



================================================================================

56 Psyche [Val. 103
clamp him against an immobile prothoracic horn or horns [Beebe (1944, 1947), Moron (1976), Eberhard (1 977, 1979, l987), Palmer (1 978), Otronen (1 988), Siva-Jothy (1 989) on scarabeids; Eberhard (1979) on ciids]; elongations of the mandibles that can be opened and closed to clamp the opponent [Goldsmith (1987) on a cerambycid; Bechtle (1 977) and Hamilton (1 979) on lucanids; Eberhard (1 979) on a tenebrionid]; and immobile, more or less straight horns that project anteriorly or posteriorly and are used as levers to pry opponents from the female or from the substrate [Brown and Siegfried (1983); Brown and Bartalon (1 986); and A. Pace (personal communication) on a tene- brionid; Eberhard (1981) on a chrysomelid; and E. Sleeper (personal communication) on a brenthid]. In many species the winning male physically removes the loser from the vicinity of the female (i.e., tosses him to the ground), or blocks his access to her (i.e., pushes him from the tunnel leading to her and defends against reentry). Apparently the only armed weevils whose behavior has been care- fully studied to date do not have horns. Males of Rhinostomus barbi- rostris use their thick, elongate rostrum and their long front legs to pry other males and flip them from logs where females are ovipositing (Eberhard 1983); and males of Macromerus bicinctus use their swollen front tibiae and elongate front legs as clubs to strike opponents in threat displays (Wcislo and Eberhard 1989). Fragmentary observations of the homed Centrinaspis sp. indicated that males apparently push each other (Eberhard and Gutierrez 199 1 ; W. Eberhard, unpublished), but the possible role of their horns was not determined. The only other published observation of which we are aware concerning possible use of prothoracic horns or spines in a weevil is that of Lyal (1986) of a single encounter between males of the zygopine genus Mecopus, in which "the thoracic spines did not appear to be employed, although the movements were so rapid that precise observations were not possible." There are many other types of horns whose mechanical designs suggest that they are not used in any of the ways documented for other beetles, and whose functional significance thus remains unknown. This study concerns one such design in the small (approximately 2 mm long) baridine weevil Parisoschoenus expositus (Champion 1908). Males of P. expositus have a pair of horns or spines projecting anteri- orly from the prothorax, and a deep invagination in the anterior surface of the prothorax between the bases of the horns. Large males have



================================================================================

20001 Eberhard & Garcia-C. 57
horns that project to the rostrum, while very small males completely lack horns (Fig. 2).
Collections and behavioral observations were made during the day near Parrita, Puntarenas Province, Costa Rica (elevation about 20 m) on 19-30 January 1998, in a plantation of approximately 8 m tall African oil palms (Elaeis guianensis). Beetles were observed on the cut petioles of the large (several meters long) leaves which had recently been trimmed from the trees and whose pinnae were still green. Most observations of behavior were made using a 2X headband magnifier and a 10X hand lens (working distance 2 cm). The combined magnification allowed detailed observations, and confident field evalu- ations of approximate male horn size without collecting and measuring beetles (large horns were about 0.3 mm or more; short horns were less than about 0.3 mrn). Approximately two hours of behavior was video- taped, using a Sony CCD-TR700 camcorder with +7 close-up lenses. All drawings illustrating behavior were traced from video images. Those portions of the animals' bodies that were not clear in the videos were omitted in the drawings.
Possible differences in colonization behavior were checked by mak- ing fresh cuts with a machete near the bases of trimmed leaves and set- ting them out in the same piles of trimmed leaves. Beetles were collected both on the surfaces of cuts, in the cracks that formed when the rachis dried, cracks where the leaf had been split when it was cut, and the narrow spaces between the matted tissue and fungus that were generally present on what had been the ventral surface of the leaf. Col- lections of beetles on open surfaces and in cracks were made around midday (10:OO-13:OO hrs) on leaves which had been on the ground for at least 4-6 days.
Beetles were measured using an ocular micrometer to the nearest 0.025 mm. Each specimen was aligned for measurement in lateral view (Fig. 1) by adjusting its position until the tip of one horn lay just over the tip of the other horn. The lateral curvature of horns was estimated by positioning the beetle's ventral surface upward so that the bases of the horns and their tips were all in focus at once, and then measuring the distances between the bases of the horns, between their tips, and between their midpoints.




================================================================================

Psyche
[Vol. 103
Fig. 1. Measurements made on specimens: (a) prothorax length; (b) horn length; (c) dorsal curve of horn.
Measurement precision was tested by remeasuring beetles on differ- ent days. The average difference between repeated measurements of the prothorax length of 51 specimens was 0.0083 k 0.0138 mm, or about 0.8% of the average prothoracic length; respective values for remeasurements of the horns of 17 specimens were 0.0088 å 0.015 mrn or 2.2%. Averages are given followed by * one standard deviation. Estimates of the densities of setae on horns were calculated from SEM images assuming the horn was a cylinder and that half of the horn's surface was visible in lateral views.
Specimens to be examined with the SEM (S-2360N) were dehy- drated from glutaraldehyde and Karnovsky solution, dried by sublima- tion, and coated with 20 pm of gold. Voucher specimens have been deposited in The Canadian Museum of Nature (Ottawa), The Museum of Natural History (London), the U. S. National Museum (Washington, D. C.), and the Museo de Insectos of the Universidad de Costa Rica.



================================================================================

Eberhard & Garcia-C.
A. Morphology
All but the very smallest males of Parisoschoenus expositus had a pair of rigid, pointed horns (or spines) projecting anteriorly from the ventral portion of the anterior surface of the prothorax (Fig. 2). The shapes of these horns varied substantially with male size. In small males the horns were small nubbins (Fig. 2) or were entirely lacking. The horns of medium-sized males were nearly perfectly straight, and each was directed anteriorly and somewhat laterally (Figs. 2, 3). The horns of large males curved both laterally and dorsally. The surface of horns of all sizes was smooth and relatively free of the setae that cov- ered most other body surfaces (Fig. 2). Horns bore only scattered, short setae, each set in a small pit (Figs. 2, 4). In two large males carefully examined in the SEM the density of these setae was greater near the base of the horn (about 2375 p.m21seta) than near the tip (about 4300 pm2/seta).
Larger beetles had larger horns (Fig. 5), and a statistical analysis of all males except those completely lacking horns similar to that of Eber- hard and Gutierrez (1991) (but using untransformed data, since trans- formations did not improve fits) revealed that there were two different body plans (both deviation from linearity and the test for a switch point were significant, p < 0.0 1). The percent of the explained variance in horn length increased from 74% with a single regression line to 85% with a two part regression, using the break point of horn length of 0.42 mm (Fig. 5). The distribution of horn lengths was flatter than that of prothorax lengths (Fig. 5). The distribution of female prothorax lengths was very similar in shape to that of the males. The horns of large males were more curved dorsally and laterally than those of small males, but the patterns of difference were not the same. Males with horns less than about 0.30 mrn had no dorsal curva- ture, and from this size upward the amount of curvature increased steadily with greater horn length (Fig. 6a). In contrast, even in males with short horns the horns projected laterally, and the amount of lateral curvature increased with horn size until the horns were about 0.5 mm and then leveled off (Fig. 6b).




================================================================================

60 Psyche [vol. 103
Fig. 2, Horns md pits of males: (a) sdl male with shoa siraight horns and only a hint of a pit meen them; (b) hge male with lomg curved horns and a pit between their haw; (c) antexior view of the oping of the pit ofthe lqe male, showing ita hled edges and &e low density of setae; id) close-up view of the inner surface ofthe pit.



================================================================================

Fig. 3, Lateral and dorsal views of the horns and prothoracic sheaths of a large {above) and a mcdium4zed male (below) drawing^ of sheath were made by dissecting away the prothomcic wall).
Fig. 4. Short, presumably sensory setae in small pits on the surface of the horn of a large male.




================================================================================

Psyche
Horn
60
Length (mm)
Length (mm)
. .
2
1 -** 8
a :: 5 +
@> 6
- 6 a .
8 . 0 0
* m
. .
; . .
0.75 1 .O 1 -25
Prothoracic length (mm)
Fig, 5. Relationship between prothorax length and horn length of males. The insets show the frequency distributions of prothorax and horn measurements. The surface of the prothorax of moderate and large-sized males was deeply invaginated, forming an oval pit between the bases of the horns (Figs. 2b-d). The surrounding surface had few setae and sloped toward the pit (Figs. 2b, c). Internally the pit led to a short central tube whose internal walls were also relatively smooth and lacked setae (Fig. 2d). This tube extended posteriorly and dorsally, and branched to form a pair of long, relatively straight, blunt-ended tubular arms which extended deep into the prothorax (Fig. 31, The internal diameter of each sheath was substantially greater than the diameter of the same male's horns (Fig. 3). The cuticle of each sheath was strong and inflex- ible. The branches of the sheath were straight, even in large males whose horns curved strongly (Fig. 3).




================================================================================

0 0.25 0.5 0.75
Length of horn (mm)
0.25 0.5
Length of horn (mm)
Fig. 6. Relationships between amount of dorsal (a) and lateral (b) curvature of horns and horn length.
The fimctional significance of the sheaths was revealed by placing males in positions similar to those seen during horn locking fights in nature. The near-side horn of each male entered the opposite-side sheath of the other male, occupying approximately half the inner diam- eter of the sheath (Fig. 7). Thus, when the right horn of male A entered the left sheath of male B, the right horn of male B entered the left sheath of male A. Because of the rigidity of the horns and sheaths, it was impossible to insert one male's horn into another similarly horned opponent unless the opponent's horn was simultaneously inserted into the first male's sheath. In two pairs of specimens that were manipu-



================================================================================

64 Psyche [Val. 103
lated into horn-locking positions (one pair of large males, the other of medium-sized males), the angle between the dorso-ventral axes of the two males measured, respectively, 71' and 73'. Thus in nature each male must have to tilt about 36' laterally away from an opponent to engage him in a horn-locking battle.
right horn of
male A
male A
Fig. 7. Horn-locking positions of two partially disarticulated large males. The right horn of male B (stippled) is lodged deep in the right sheath of male A. B. Behavior
Beetles occurred on both fallen trimmed leaves and the cut leaf stub remaining on the tree. Beetles that had wedged themselves into cracks on or near the cut surface (and into folds of plastic bags where some were kept temporarily) were immobile. This immobility and the lack of space in which they could execute the maneuvers associated with mat- ing, fighting and oviposition suggest that their sexual behavior does not occur in cracks. Beetles found during the morning on less 4ecently cut leaves, whose pinnae were still green but beginning to brown, were all in cracks. No beetles were found on leaves whose rachis was still green but whose pinnae had all turned brown.
1. Oviposition
Drilling and oviposition were preceded by apparent searching behavior. The female walked slowly over the cut surface of the leaf, repeatedly touching it with the tip of her rostrum and her antennae. She often paused with the tip of her rostrum touching the palm for variable amounts of time that lasted up to several minutes. It was not clear whether these pauses represented feeding or searching for oviposition



================================================================================

Eberhard & Garcia-C.
sites. Of a total of 22 females that were followed after an oviposition, 9 eventually laid a second egg, 9 walked into a crack or off the cut sur- face? and 4 more were lost from view (probably also went off the cut surface). None of these females laid a third egg while being followed. The time between ovipositions ranged from 2 to 43 min. (mean 19.3 min.).
Oviposition was always preceded by a period of "drilling," during which the female ceased walking and inserted her rostrum into the leaf (e.g., Fig. 9). The shortest drilling was just under 2 min.? while the longest was over 6 min. The female sometimes inserted her rostrum all the way to her eyes. In at least one case a female made a deep hole? but then abandoned it without ovipositing. It is possible that some "drilling" was actually feeding behavior. There were few if any percep- tible prying or turning movements during drilling, such as were per- formed by drilling Metamasius sp. females on the same palm leaves. Nor did drilling females bring pieces of plant tissue ("sawdust") to the surface or work at the edge of the hole just before ovip~sition~ as do those of Rhinostomus barbirostris (Eberhard 1983). The end of drilling was marked by a smooth withdrawal of the rostrum and a 180' turn. As will become clear, the lack of overt signs that distinguish female searching or feeding behavior from drilling and impending oviposition has important consequences for understanding male behavior. When the female turned 180' after drilling? she positioned the tip of her abdomen at the mouth of the hole left by her rostrum, sometimes after brief searching behavior with the tip. The tip of her abdomen was extended, and she remained immobile for an average of 35 k 10 sec- onds (N = 24). She then immediately turned 180' and "worked" on the material where oviposition had occurred with the tip of her rostrum for an average of 72 12 seconds (N = 5). Both her antennae repeatedly touched the surface of the palm near the tip of her rostrum, and the tip of the rostrum appeared to pull together and gently tamp down material on the surface.
2. Courtship
The male mounted the female with no preliminary courtship, and positioned himself to face in the same direction. Females were almost never mounted if they were walking. After a variable amount of time (ranging from <lo seconds to 166 seconds in 18 pairs), the male per- formed courtship behavior and then attempted intromission. The first



================================================================================

66 Psyche [vole 103
behavior pattern (omitted in some pairs that nevertheless copulated) was to briskly rub the tip of his rostrum back and forth across the sur- face of the female's pronotum. The exact direction of these rubs varied? but was often largely from side to side. Very small males mounted on large females rubbed the female's elytra rather than her pronotum. Following rostrum rubbing? the male moved posteriorly on the female to position the tip of his abdomen near the tip of hers, and vig- orously rubbed the lateral and ventral surfaces of the posterior portion of her abdomen with the ventral surfaces of his folded hind tibiae (Fig. 8a). Tibia1 rubbing occurred in bursts of approximately a second, and was either accompanied or immediately followed by partial eversion of the male's genitalic basal lobe and one or a series of small? rapid stabbing movements with it against the tip of the female's abdomen. Intromission occurred at this stage if the female was receptive and opened the tip of her abdomen. Rubbing was sometimes reduced to as little as a single rub, and sometimes omitted entirely. If Fig. 8. Patting behavior: (a) a male raises both fiont legs (stippled) preparatory to pat- ting the female during copulation; (b) a partially dismounted male pats the female soon after copulation ended (dotted lines followed others by 0.1 second).



================================================================================

intromission was not successfil, the male either rubbed again with his hind tibiae and tried to intromit again (often success~lly), or moved forward again. The maximum number of intromission failures was three, and most males were successfi.d on their first or second intromis- sion attempts. One male that had just failed to intromit moved forward? vibrated his entire body briefly (see description of stridulation in the section on aggression below), and moved back to rub with the hind tib- iae and attempt intromission again. In no case was there any sign of female attempts to dislodge males by kicking or other body move- ments. Nor was there any sign of contact with or other display of the male's horns.
3. Copulation and after
Copulation lasted an average of 100 k 46 seconds (N = 29). The male was usually immobile during most of copulation, except for a steady pumping movement of the tip of the male abdomen [Fig. 9a; see Eberhard (1994) for a verbal description of very similar movements in seven other species of curculionids]. When another beetle passed nearby? however, the male usually vibrated his head rapidly dorso-ven- trally (Fig. 9b). This movement appeared to be designed to produce sound by rubbing the head against the anterior edge of the prothorax, but there were no stridulatory structures on the head or prothorax, and female morphology in this region was similar to that of males. Head movements may instead finction as visual signals. Two males rubbed the female's pronotum briefly during copulation with the rostrum as in pre-copulatory courtship behavior. In 5 of 13 carefilly observed pairs? the male became active during the last approximately 5-15 seconds of copulation, patting the female rapidly on the pronotum with one or both of his front legs (probably contacting her with the tips of his tibiae and tarsi; see Fig. 8b).
The male ended copulation by withdrawing his genitalia and step- ping posteriorly, often dismounting to stand just behind the female with his head and at least part of his prothorax over her elytra. Usually (16 of 19 cases checked for this detail) the male patted the female's dorsal surface with one or both front legs in one or several shorts bursts of movement after dismounting (Fig. 8b). In the most elaborate form of these post-copulatory displays the male stood with both fiont legs raised and partially extended anteriorly, and then delivered several bursts of rapid simultaneous pats with both front legs. Each pat lasted



================================================================================

Psyche
Fig. 9. Movements of the male during copulation: (a) pumping movement of the tip of the male's abdomen during copulation (dotted lines follow others by 0.1 second); (b) a large male vibrates his head dorso-ventrally while copulating with a drilling female (dot- ted lines follow solid by 0.07 second).
about 0.03 second, and bursts of pats lasted up to a second or so. As in pre-copulatory courtship, the male horns were never in positions that would allow them to stimulate the female tactilely; nor was the male positioned appropriately to emphasize their visual impact on the female.
A drilling female often copulated with several males, and the final copulation frequently ended only shortly before oviposition. In 22 pairs, oviposition followed the end of copulation by <30 seconds in 8, and by <60 seconds in 14. Male defense of the female (next section) usually ended abruptly as soon as the female turned to oviposit. In 14 of 24 carehlly timed pairs the male left within 10 seconds after the female turned (in some pairs he departed less than 1 second after she turned). In only 1 of the 24 was the defending male still near the female 60 seconds after oviposition began. Of 58 cases in which the initiation of oviposition was observed, the last male to accompany the female was judged to be large-horned in 48 (in 3 1 of these cases copu-



================================================================================

20001 Eberhard & Garcia-C. 69
lation by this male had been observed); and in 9 the last male was judged to be small-homed (in all 9 pairs copulation by this male had been observed) (the other case involved a male whose horns were intermediate in length).
Females never showed any resistance, such as walking, kicking, or shaking their bodies, to being mounted or ridden by a male. Many mounts, however, did not result in copulation attempts because the male dismounted immediately and moved away if the female was not drilling, especially when she began to walk (in 12 of 15 cases checked for this detail). Thus a female in the process of searching for an ovipo- sition site was generally mounted briefly but then abandoned by a series of different males (and sometimes repeatedly by the same male). The longest such fruitless mount, which involved a female that was not walking, lasted 1 0- 1 1 min.
A second common type of fruitless mount was while the female was motionless during oviposition, during tamping behavior following oviposition, or when her abdomen protruded while she was motionless with her anterior end inserted into a small crack. Males almost always abandoned such females immediately, however, as soon as they began to walk.
4. Aggression
Females never interacted aggressively with each other, except in a single case in which one female attempted to push her rostrum under that of another which was drilling. In contrast, male-male aggression was very common. Afier a male copulated with a drilling female, he always remained with her and attempted to defend her against other males. Aggressive behavior in defense of the female took several forms that will be described in order of increasing intensity. Usually a male's first defensive response to another beetle approaching fiom in front of the female was to move forward onto the female and vibrate his head dorso-ventrally (Fig. 9b) and make patting movements with his front legs. Patting in this context amounted to hitting movements directed toward the other beetle that fell short. If another beetle approached from the side or the rear, the male leaned toward the intruder, thus interposing his body between the other beetle and the female. Sometimes this maneuver succeeded in keeping the other male (even a larger one) from encountering the female, and he walked on without hrther interaction. In nearly all cases in which a large male