ONTOGENY OF DISPLAY IN IMMATURE SCHIZOCOSA CRASSIPE8 (ARANEAE: LYCOSIDAE)

Most studies dealing broadly with behavioral ontogeny in spiders have been concerned with orb-weaving (Witt and Baum, I96O; LeGuelte, I969; Peters, I969; Reed, Witt, Scarboro and Peakall, 97o). However, relatively little attention has been directed to the ontogeny of display. Immature lycosid spiderlings do not exhibit the complex stereotyped displays typical of adult wolf spiders during courtship (Rovner, I968) or agonistic encounters (Aspey, 1974, 1975). Crane (1949) reported th,a.t several species of immature salticid spiders occasionally performed abortive displays, but she considered such displays exceptions to the rule. Rovner (1968) looked unsuccessfully for courtship display in penultimate male Lycosa rabida; the tendency to court did not appear until several days after the final molt. The first study of display ontogeny in spiders was that of Dijkstra (1968, in Koomans et al., 1974), which was expanded by Koomans, van de’r Ploeg and Dijkstra (1974) who .observed a characteristic "leg wave behavior" in a variety of subadult Pardosa spp. This behavior vas observed from the second instar to the adult molt in P. lugubris and P. nigriceps, and decreased in frequency with concomitant increases, in courtship. In considering the functional significance of leg wave behavior, Koomans et a.1. (1974) suggested that courtship replaced the behavior in adult males, while in adult females, the behavior was probably replaced by specific reacti.ons to. courtship. Dijks.tra (1968, in Koomans et al., I974) observed similar leg movements in juvenile P. amentata, but the behavior was not restricted to courtship situations. Van der Ploeg (personal communication) has observed leg waving in immature P. purbeckensis in the field. Similarly, I had observed (Aspey, I974) that both male and female Schizocosa crassipes (Walckenaer) exhibit a characteristic leg wave display as juveniles.

97o). However, relatively little attention has been directed to the ontogeny of display. Immature lycosid spiderlings do not exhibit the complex stereotyped displays typical of adult wolf spiders during courtship (Rovner,I968) or agonistic encounters (Aspey, 1974(Aspey, , 1975. Crane (1949) reported th, a.t several species of immature salticid spiders occasionally performed abortive displays, but she considered such displays exceptions to the rule. Rovner (1968) looked unsuccessfully for courtship display in penultimate male Lycosa rabida; the tendency to court did not appear until several days after the final molt. The first study of display ontogeny in spiders was that of Dijkstra (1968, in Koomans et al., 1974), which was expanded by Koomans, van de'r Ploeg and Dijkstra (1974) who .observed a characteristic "leg wave behavior" in a variety of subadult Pardosa spp. This behavior vas observed from the second instar to the adult molt in P. lugubris and P. nigriceps, and decreased in frequency with concomitant increases, in courtship.
In considering the functional significance of leg wave behavior, Koomans et a.1. (1974) suggested that courtship replaced the behavior in adult males, while in adult females, the behavior was probably replaced by specific reacti.ons to. courtship. Dijks.tra (1968, in Koomans et al., I974) observed similar leg movements in juvenile P. amentata, but the behavior was not restricted to courtship situations. Van der Ploeg (personal communication) has observed leg waving in immature P. purbeckensis in the field. Similarly, I had observed (Aspey, I974) that both male and female Schizocosa crassipes (Walckenaer) exhibit a characteristic leg wave display as juveniles. In extending the generality of leg waving display among immature lycosids, this study describes the ontogeny of leg wave display anaong immature S. crassipes and suggests a biological signi/icance for the behavior.

METHODS
The subj'ects were 64 immature S. crassies. Thirty-two spiders (I6 males ,and I6 females) were collected during early May, I97-73, as antepen-or penultimate instars from Stroud's Run State Park, Athens Co., Ohio, USA. Th'e remaining spiderlings were reared from laboratory-mated adults. These spiderlings were observed from egg case emerge,nee (.considered the second instar) until approximately wk after dispersal from the female. For most spideriings, one additional molt (to the third instar) occurred in the laboratory, but too few animals survived beyond this time to continue observations. However, observatio,ns continued until the qnal molt for those spiders collected as antepen-and penultimate instars in the field. Data are reported only for those individuals for whom sex was correctly established at the antepen-or penultimate instars.
Six antepen-and penultimate spiderlings of each sex collected from the natural habitat and I2 laboratory-maintained spiderlings were isolated in covered clear plastic containers (I2.5 X 7.o X 7.o era). One-hal of these spiders were visually isolated, while the remaining containers were adjacent and allowed the, spiders visual access to one another. Ten antepen-and penultimate spiderlings of each sex collected from the field were socially housed (sexes separate) in glass terra.ria (26 X I9 X 22 cm) fitted with plastic covers. Two groups of laboratory-maintained spiderlings (IO per group) were observed during their second and third instars. In order to observe these animals under the above conditions, spiderlings were brushed off the female's abdomen with a camel hair brush. Ten individuals were then placed in an observation chamber identical to the one in which socially grouped antepen-and penultimate instars were maintained. At the conclusion of each observation period the spiderlings were returned to the f.emale and allowed to regroup. After dispersal from the female began, two groups of O individuals were transferred to separate observation chambers until the third instar, at which age observations were terminated.
The frequency (i.e., bouts per 30 rain) and duration of leg waving were recorded for 30 rain daily with a manually activated Ester- Vestigial-winged Drosophila melanogaster were of-fere'd as. food to each spiderling once weekly, and a constant water .supply was available. The spiderlings were. maintained and observed under relatively const,ant temperature (22-26C) and humidity conditions (55-62% RH).

RESULTS
Leg waving display in immature S. crassipes consisted o the simultaneous raising and lowering o the /]rst and second ipsilateral legs. Leg waving was generally p.erormed at a steady and smooth rate, and not alternated rom side to side during a leg waving bout. Spiderlings showed no preference or either the right or let side during leg waving. Although leg waving occurred nearly as oten per 30 min during all instars, no consistent temporal patterning was evident until the antepen-and penultimate instars, at which time leg waving became increasingly stereotyped and sexually differentiated.
Leg vaving was observed only /]ve. times in the 24 socially isolated spiderlings. Four o these instances were limited to those spiders having visual access to one another, while, only one visually isolated spiderling exhibited leg waving. The five observations o leg waving occurred in emales once each at the second, third, and antepenultimate instars, and in males twice during the penultimate instar.
Although leg waving was rarely observed in socially isolated spiderlings rega.rdless of age or sex, socially grouped spiderlings o both sexes at all the developmental stages studied exhibited leg waving.
Of the IO male.s and IO females collected from the natural habitat and grouped separately by sex, nine males and our emales reached sexual maturity. Cannibalism among these spiders was directly observed on two occasions.; the remaining deaths may have been due to other causes. In the two groups of IO laboratory-maintained juveniles, two individuals from each group reached the .ourth instar.
Among these spiders cannibalism was also, directly o.bserved on two occasions. Among socially gro.uped antepen-and penulatimate males, bouts o leg waving lasted an average o 7.3 +/-2.1 (SD) sec, and consisted of the legs being waved 3-27 times ( +/-SD" 16.o __+ 4.7) in succession at a rate o.f approximately 2 waves/sec. With increasing age, ilm analysis revealed that the foreleg was lifted higher off the substratum (20-25 in the second and third instars) until an rc 40-5 0 relativ.e to the substratum was reached in the antepenand penultimate instars. The second ipsilateral leg that accompanied the foreleg during leg waving was lifted only 5-25 relative, to. the substratum. The legs were raised and lowered almost simultaneously, with the foreleg occasionally leading.
Among socially grouped antepen-and penultimate females, bouts of leg waving lasted an average o.f 4.I -+-2.0 sec, and consisted of the legs being waved -IO times (ff -+-SD" 3.o -+-2.2) in suecession .at a rate o:f approximately wave/sec. Although emales exhibited leg waving more slowly and less trequently than males, the behavioral to,pology of leg wa.ving was indistinguishable between the sexes. Bouts of leg waving alternated with variable periods of immobility or locomotion, and depended o.n the behavior of conspecifics.
The circumstances under which leg waving occurred did not seem to change with age. A spiderling initiated leg waving in response to a darting front approach by a conspecific when it approached to within two or three body lengths. In response to one or several bouts of leg waving, the approaching spider turned away and/or retreated. On the four occasions when cannibalism was observed, the cannibalized spiderling was approached from the posterior; leg waving was not observed to occur in response, to these, posterior approaches. o.f the Pardosa species, the pattern of leg wave behavior was gen-,erally the same at all deve.lopmental stages, with the frequency decreasing rapidly at the adult molt. In S. crasipes leg waving patterns become increasingly stereotyped with age, but the frequency did not vary with age or decrease as the final molt neared. Blinded Pardosa showed relatively little leg wave behavior, similar to. the lack of leg waving ,exhibit:ed by socially isolated S. crassipes. Although both sexes of Pardosa spp. and S. crassipes exhibited leg waving, antepen-and penultimate male S. crassipes displayed significantly more than females. Of further interest were the findings that bouts of leg waving by males lasted longer and consisted of more waves than those exhibited by females. These findings, coupled with the observation that more males than females reached sexual maturity among socially grouped spiderlings, suggested that leg waving serves a communication unction that becomes more sexually differentiated with age.. This idea is consistent with Aspey's (in preparation "a," "b," "c") research that only adult males exhibit the highly complex agonistic display which serves to. maintain dominance-subordinance relations and preserve inter-individual distances.
The decrease in leg wave behavior concurrent with increased courtship by adult male P. lugubris and P. nigriceps suggested to Koomans et al. (I974) that the behavior was replaced by courtship.
Analogously, le.g wave behavior in adult females could be replaced by specific reactions to courtship. In immature S. crassipes leg waving may be homologous with the similar ipsilateral Prolonged Wave in adult males (Aspey, 974, in preparation "a"). Although Prolonged Wave occurred infrequently in adults relative to the requency o immature leg waving among socially grouped spiderllngs, both behaviors occurred under similar circumstances (i.e., one spider approaching another (rom the ront). These. behaviors seemed to have similar behavioral consequences, that of increasing distance between two individuals and thwarting approaching conspeciiqcs. demonstrated in jumping spiders by Drees (I952), a similar inhibition of agonistic responses may also be the role of leg wave behavior.
Leg waving probably is a mechanism to space spiders living in dense populations; this would minimize cannibalism and competition i'or food supplies. Observations on immatures and adults .support this idea. The only occasions during which observed cannibalism occurred among immatures was during a posterior approach when leg waving did not occur. The probability of cannibaiism is likely minimized during a face-to.-face encounter when the approached spider exhibits leg waving, and the approaching spider subsequently retreats.
Although immature S. crassipes exhibit the various locomotory and contact behaviors o.f reproductively mature adults (Aspey, I974, in preparation "a"), adult-like oreleg movements and postures are not observed until sexual maturity. The complex, stereotyped agonistic display o. adult male 8. crassipes serves to determine domin'ancesubordinance relations (Aspey, in preparation "b"), as well as space the animals according to a specific inter-individual personal distance (Aspey, in preparation "c"). The prominent tibial brushes of adult males, coupled with foreleg movements, probably serve as visual signals to indicate the presence of male conspe.cifics. However, with immatures, tibial brushes are lacking, and leg waving is probably the only visual signal available; it may serve a generalized spacing function for immatures o? both sexes until the various adult behaviors develop. Thus immature leg waving may be considered a mechanism to space spiderlings, with the added consequence that cannibalism is minimized and spiders are selected affainst that do not exhibit and/or respond appropriately to leg waving. ACKNOWLEDGMENTS I thank Dr. Jerome S. Rovner, Department of Zoology and Microbiology, Ohio University, o.r his helpful guidance and encourageme'nt throughout this work, and for critically reviewing the manuscript. Drs. S.W.F. van der Ploeg and H. Dijkstra, Free University, Amsterdam, The Netherlands, critically read the manuscript, and I :extend my appreciation to 'them. LITERATURE CITED zaXSPEY, W. P. 1974. Wolf spider sociobiology: An ethological and informational theory analysis of agonistic behavior in Schiocosa crassipes.