Insectes Sociaux

, Volume 57, Issue 3, pp 323–332

Synchronized contractive movement of Amaurobius ferox spiderlings

Authors

    • Division of Life Sciences, College of Natural SciencesUniversity of Incheon
Research Article

DOI: 10.1007/s00040-010-0087-0

Cite this article as:
Kim, K.W. Insect. Soc. (2010) 57: 323. doi:10.1007/s00040-010-0087-0

Abstract

During the post-matriphagy period, Amaurobius ferox spiderlings (Araneae, Amaurobiidae) show synchronous movement, contracting their bodies simultaneously. This paper describes this behavior for the first time and identifies influencing factors. The spiderlings’ contractions triggered by web vibration caused by intruders result in a strong pulsation of the whole web that a single individual would not be able to induce by itself. Repetition of the contractions was synchronized among individuals (n = 60 clutches). The movement appeared on the first day after matriphagy. The proportion of participants was maximum on the third day post-matriphagy, when on average 60.7% of the individuals were involved; thereafter the synchronicity progressively decreased. The spiderling groups performed contractions at the highest frequency on the fourth day post-matriphagy, and stopped contracting after the second molt. Experiments using mechanical stimuli produced by an electronic vibrator and a cricket’s movement showed that the vibrational intensity of the external stimuli was positively correlated with the number of contractions performed. Nestmate presence increased the number of contractions performed by individuals, and members of densely packed groups showed more contractions per individual than those in less dense groups. Contractions appeared only during the period when the mother was absent (after matriphagy, or when the mother was removed after the first molt of spiderlings and before matriphagy), and the young were not yet capable of capturing prey. Contractions may function as an antipredatory behavior.

Keywords

Collective behaviorSynchronized movementSpiderlingsAmaurobius

Introduction

Synchronized behaviors of animals, such as the flashing of fireflies, claw waving courtship of fiddler crabs, and chorusing in anurans or insects, require some form of coordination among individual behaviors with respect to time. Synchronization of behaviors does not necessarily emerge in highly social groups. Research attempting to clarify the mechanisms enabling coordination between individuals in social animal colonies is in progress, but little work has been done to date on spiders (Saffre et al., 1997; Vakanas and Krafft, 2001; Bourjot et al., 2003). Nevertheless, knowledge about the origins of these mechanisms is necessary to understand social evolution in arthropod societies (Vakanas and Krafft, 2001).

Despite the reputation of spiders as solitary carnivores, collective activities have been observed in social species (Kullmann, 1972; Lubin, 1974; Buskirk, 1981; Krafft and Pasquet, 1991; Avilés, 1997; Mailleux et al., 2008). For example, groups of Anelosimus eximius (Araneae, Theridiidae), a social spider of South America, display synchronized prey-capturing behavior (Krafft and Pasquet, 1991; Vakanas and Krafft, 2001). Krafft and Pasquet (1991) analyzed the movement of An. eximius during the hunting phase. After a prey animal falls into the web, a number of An. eximius individuals, which were initially spread out over the communal nest, approach the epicenter of vibrations sent out by struggling prey. While approaching the prey, each spider displays alternating periods of immobility and movement and after 2 or 3 s, these periods become synchronized among the approaching spiders. Once synchronized, all of the spiders in the hunting group remain immobile during the resting periods, and 70–90% of the spiders move during the active periods. This coordinated activity is supposed to increase the efficiency of their predation strategy.

Coordinative behaviors between individuals are not restricted to social spiders but occur in solitary spiders, although it has been rarely studied (Jeanson et al., 2004; Kim et al., 2005a, b). Although agonistic interactions and cannibalism occur among individuals, solitary spiders undergo a gregarious phase during their very early period of life (Krafft, 1979; Horel et al., 1996). This study describes a synchronization in body movements observed in a group of juvenile spiders. This is the first study examining the synchronization of movements in non-social spiders.

Amaurobius ferox (Araneae, Amaurobiidae) shows a similar collective phenomenon to the synchronized movement observed in the social spider An. eximius (Krafft and Pasquet, 1991). Am. ferox females produce a single clutch containing 60–130 spiderlings and exhibit maternal provisioning behavior in the form of trophic eggs (Kim and Roland, 2000) and matriphagy (Kim and Horel, 1998). After matriphagy, the Am. ferox siblings form a very compact group and remain within the natal nest. In the absence of the mother, the young form a social group for 3–4 weeks until the time of dispersal from the maternal nest (Kim, 2000). During this post-matriphagy period, the spiderlings show synchronous movements in which multiple individuals simultaneously contract their bodies to produce large-scale movements of the web.

The collective behavior of Am. ferox appears to result from the majority of the spiderlings simultaneously seizing web silk and pulling it abruptly with their legs, and then releasing it at once. As a result of this behavior, the whole web contracts and expands. The young repeat the contractive movements many times in close succession once it starts. In this paper, I call the synchronized behavior of Am. ferox spiderlings ‘contraction’.

I investigated the characteristics and dynamics of the phenomenon, including the proximate triggers for the behavior, the development and synchronization of the behavior, the influence of the siblings’ presence, and the effects of variation in group density.

Materials and methods

Natural history of the species

Am. ferox Walckenaer (Araneae, Amaurobiidae) is a subsocial spider commonly found in Europe (Cloudsley-Thompson, 1955; Bristowe, 1958). The females spin irregular cribellate webs under stones in the forest litter, in holes in old walls, and in other sheltered places (Cloudsley-Thompson, 1955; Tahiri et al., 1989). Maternal activity of Am. ferox occurs in early summer and is characterized by caring behaviors performed over a fairly constant interval from the initiation of egg-laying (Kim and Horel, 1998; Kim and Roland, 2000). With maternal help, spiderlings emerge from the egg sac after 20 days of incubation (Kim, 2009). One or 2 days after emergence, the mother provides her young with a batch of trophic eggs, which are immediately devoured (Kim and Roland, 2000). The young molt 3–4 days later (Kim, 2001) and matriphagy occurs 1 or 2 days after this first post-emergence molt (Kim and Horel, 1998). In this semelparous species, the mothers are always devoured by their young (Kim et al., 2000).

After their mother’s death, the brood forms a temporary social group (Kullmann, 1972). The siblings remain in the natal nest for 3–4 weeks until dispersal (Kim, 2000). During this period, they pass through their second and third molts, and collective prey capturing appears (Kim et al., 2005a, b).

Behavioral observations

This study was carried out under laboratory conditions because the preferred locations of Am. ferox preclude detailed observation in the field, and laboratory conditions permitted manipulation of clutches. Females were collected from early May to early June, i.e., before the egg-laying period, from under fallen stones and in ruined walls in the forested area of Nancy, France (48°41′N, 6°13′E, elevation: 217 m, annual temperature: 9.6 ± 6.3°C, annual precipitation: 74 cm).

Following collection, the females were transferred to a closed room lit by fluorescent tubes (approx. 100 lx, 12 h light/12 h darkness), and maintained at a temperature of 20 ± 2°C. Each female was placed in an individual glass terrarium (L 200 mm, W 120 mm, H 200 mm) partly filled with a mixture of sand and peat (Tahiri et al., 1989; Gundermann et al., 1993), which was humidified twice a week. The spiders were liberally provided with cricket nymphs (Gryllus bimaculatus). Most females (>90%) oviposited. Laboratory experiments were conducted during the years of 1995–1997 at the Laboratoire de Biologie et Physiologie du Comportement, Université Henri Poincaré, Nancy 1, France.

I observed the behavior of spiderlings in a total of 60 clutches on the third day post-matriphagy. I introduced animals that had been frequently observed in the natural habitat of Am. ferox (each species of animal in ten clutches). Introduced animals included cricket nymphs (20–40 mg; Orthoptera), ants (10–20 mg; Hymenoptera), earthworms (120–180 mg; Oligochaeta), wood lice (30–50 mg; Isopoda) and wasps (50–60 mg; Hymenoptera), as well as other species of spiders frequently observed in the habitat of Am. ferox: Amaurobius fenestralis (70–100 mg; Amaurobiidae), Coelotes terrestris (150–200 mg; Agelenidae), and Tegenaria atrica (250–350 mg; Agelenidae). Body sizes of the introduced animals above were all much larger than that of spiderlings (body mass: 2.0–2.6 mg, body length: 2.6–2.8 mm). I also introduced mites (Acarina), which are often observed in the internal area of the web and smaller (<1 mm) than a spiderling. After introduction of each species, I observed ten clutches for 10 min and recorded whether the clutch showed collective movements or not.

Synchronous properties of the behavior

Motion analysis software (MAS)

In order to show synchronization of the behavior among individuals, I used MAS (4th ed., Keio Electronic Industrial Co. Ltd., Osaka, Japan) which analyzes video images to measure the velocity of identical objects. The software digitizes images, representing individuals as black spots on the light background, and allows to follow ten individuals simultaneously. The digitized images did not include all individuals within the group. However, the measurements provide an alternative method to quantify and analyze the behavior.

Proportion of individuals involved in synchronized movement

I observed individual behavior in contractions on video-recorded images (n = 20 clutches). For this experiment, I introduced a cricket nymph onto the web to induce the spiderlings’ movements. After matriphagy, the mean body mass of a single spiderling was 2.3 ± 0.2 mg (measurement of five individuals in each of 12 clutches), whereas the mass of the cricket nymph was 20 ± 2 mg. A cricket nymph of this size could, therefore, be a potential predator for individual spiderlings or alternatively, a prey item for the whole group (see Kim et al., 2005a).

The behavior of the spiderlings was videotaped for 10 min after introduction of the cricket, during which I estimated the proportion of individuals in the group participating in each synchronous contraction. I repeated the experiment and recorded changes in the number of individuals participating in synchronous contractions on the third, fourth, fifth, and sixth day post-matriphagy.

Ontogeny of the behavior

In order to investigate the behavioral development (in relation to the age of the spiders) of the phenomenon during the post-matriphagy period, I measured the frequency of collective contractions everyday from the first day to the seventh day post-matriphagy in ten clutches that remained on the maternal web in the glass terrarium throughout the experiments.

Factors influencing behavior

I conducted experiments to investigate the proximate mechanism involved in the phenomenon. I examined external stimuli that trigger the behavior of the spiderlings, the effects on the behavior of the intensity of the external stimuli, the number of individuals in a group, group density (distance between individuals) and the quantity of silk in the web.

External stimuli

The collective movement of the spiderlings might be a response to vibrational stimuli transmitted via the web, or a response to chemical stimuli possibly radiating from a stranger. In the absence of vibrational stimuli, the spiderlings showed little locomotor activity and did not perform contractions. I tested whether an artificial vibration is sufficient to elicit the behavioral response of spiderlings. I also examined effects of intensity of the vibration in this experiment.

The electronic vibrator with an iron skewer was set to produce 0.5 vibrations per second with a square wave at three different intensities (amplitude): strong (1.5 cm), medium (0.7 cm), and weak (0.3 cm). The amplitude of electronic vibrator was arbitrarily decided from web shakes caused by movements of a 20-mg cricket tangled into the web of Am. ferox: the amplitude of strong vibration was based on the shaking caused by the cricket’s whole body movement, and the weak vibration based on minimal movement of the cricket (see below).

Fifty spiderlings from each clutch (n = 5) were moved to a transparent box (L 160 mm, W 90 mm, H 75 mm) on the first day after matriphagy. The transparent box was custom-designed for close-range video recording and observation by direct sight and stereomicroscopy (Zeiss operation). To promote web construction, one corner of the box was darkened by covering it with black paper. A transparent film covered this corner to prevent the spiderlings from spinning against the lid of the box. A piece of cotton placed in a corner of the box was humidified and the lid of the box was pierced with a grid-covered window for aeration (Kim, 2001).

I conducted the test on the fifth day, when the spiderlings wove a communal web while remaining in a group. The clutch was exposed to vibrations of each of the three amplitudes (strong, medium, and weak) for 3 min at intervals of 10 min. The order of presentation of the three vibrations was randomized for each clutch. Following the activation of the vibrator, I observed the number of contractions performed by the spiderlings.

Second, I used a 20 ± 2 mg cricket (G. bimaculatus) nymph to generate natural vibrations. I conducted this experiment on 20 clutches on their maternal webs in the terrarium on the third, fourth, fifth, and sixth days post-matriphagy and observed the responses of the individuals to the movements of the cricket.

Movements of the cricket were classified into three categories: (1) ‘body’ when the cricket was moving its whole body, which created strong vibrations on the web; (2) ‘appendage’ when only a part of the body (a leg, an antenna or the head) was moving; (3) ‘absent’ in the absence of movement (when the cricket was present on the web).

On the 10-min video sequence recorded immediately after the introduction of the cricket onto the web, I measured the number of contractions in which each individual participated. I then counted the number of contractions in which an individual participated per 100 s in response to each of the three types of movement by the cricket.

Presence of nestmates and the size of the web (quantity of silk)

As the behavior of spiderlings is a collective phenomenon, the absence or presence of nestmates might influence the behavior, and as the vibrations that trigger the contractions are transmitted by the silk structure of the web, I also predicted that there would be an effect of the quantity of silk on the behavior of the spiderlings. To examine the effects of these variables, I conducted experiments under three treatments:
  • Treatment 1: ten individuals within a web woven by ten individuals (n = 13).

  • Treatment 2: one individual within a web woven by ten individuals (n = 16).

  • Treatment 3: one individual within a web woven by one individual (n = 15).

The spiders were moved from the maternal nest on the first day post-matriphagy and installed in a transparent box (L 100 mm, W 70 mm, H 75 mm). The lid of the box was made with a grid-covered window for aeration and a piece of cotton placed in a corner of the box was humidified.

For treatment 1, ten siblings that had been installed in a plastic box and had constructed a communal web were taken off the web on the fifth day. I used CO2 to remove the spiders so as not to damage the web. Subsequently, I replaced the ten individuals on the empty web. For treatment 2, only one individual taken from the web was replaced on the empty web woven by the ten individuals. And in the treatment 3, a single individual was removed and returned to its web.

On the fifth day post-matriphagy, a video recording was made for 3 min following the introduction of a cricket. In this experiment, in which a small number (1 or 10) of individuals was obliged to construct a new communal web, I placed a 5 ± 0.5 mg cricket nymph on the web. Individual movements were analyzed on the video recording. The durations of the cricket’s movements were statistically equal in the three treatments (ANOVA: F2,41 = 0.077, P = 0.926). I then compared the number of contractions produced in the three treatments.

Group density

The behavior of one individual might influence other individuals’ behavior in compact groups of spiderlings. Moreover, information can be transmitted between individuals by the web silk. Therefore, it is possible that the distance between individuals (i.e., the local density) has an influence on the collective movement. The majority of the clutches formed a compact group during the experimental period. However, some clutches were more scattered on the web than others, permitting me to examine the relationship between local density and collective movements.

Video recording was conducted on 60 clutches for 10 min following the introduction of a 20 ± 2 mg cricket on the fifth day post-matriphagy. I then analyzed the number of contractions performed by each individual in each clutch for 10 min.

On the video image I randomly selected five individuals per clutch, and measured the distance to the closest individual relative to individual body length. Then, I divided the groups into two categories: ‘dense’ and ‘less dense’. ‘Dense’ groups were those in which the average distance between individuals on the video image was less than the length of an individual, and ‘less dense’ groups were those in which the average distance between individuals was greater than the individual body length.

Spiderlings in the same clutch have almost equal body lengths at this age. This developmental evenness among individuals within a clutch results from matriphagy, as all spiderlings within clutches benefit from the intake of their mother’s body (Kim et al., 2000). Matriphagy results in a 2.5-fold weight gain in the young over their initial weight (Kim et al., 2000).

Of the 60 clutches examined, 46 clutches were categorized as ‘dense’, and 13 clutches were ‘less dense’ (the mean interindividual distance for these clutches was approximately twice the body length of an individual). The number of individuals within the clutches was not statistically different for the two density categories (dense: 77.7 ± 9.2, less dense: 75.9 ± 13.2; Mann–Whitney U test: n1 = 46, n2 = 13, z = −1.189, P = 0.235). The time (duration) of the cricket’s movements was also not statistically different between the two density categories (Mann–Whitney U test: n1 = 46, n2 = 13, z = −0.786, P = 0.432).

Statistical analysis

Statistical analyses were performed with StatView (version 5; SAS Institute 1998). Differences among individuals participating in each contraction were compared using Chi-square test of independence. The proportion of individuals within a clutch participating in a contraction was compared among the days post-matriphagy using repeated measures ANOVA. Changes in the proportion of individuals participating over time were examined using ANOVA for linear regression analysis. The numbers of contractions observed in response to the different intensities of vibrations produced by an electronic vibrator and to different types of movements by the cricket were compared using Friedman test for non-parametric repeated measurements. The number of contractions produced in three different social and structural treatments was compared using ANOVA and comparison between groups of different densities was carried out using the Mann–Whitney U test. The total duration of the cricket’s movements was compared across experimental treatments by ANOVA. Comparisons of clutch sizes and times that the crickets spent moving between clutches with different densities were carried out by Mann–Whitney U test.

Results

Description of the behavior

Contractions were performed by all the observed clutches (n = 60). On the third day post-matriphagy, the spiderlings showed the synchronous movement when I introduced animals frequently found in the habitat of Am. ferox to the web: crickets (Orthoptera) (n = 10 clutches), ants (Hymenoptera) (n = 10), earthworms (Oligochaeta) (n = 10), wood lice (Isopoda) (n = 10) and wasps (Hymenoptera) (n = 10). The spiderlings also produced the movement in the presence of the sympatric spider species Amaurobius fenestralis (Amaurobiidae) (n = 10), C. terrestris (Agelenidae) (n = 10), and T. atrica (Agelenidae) (n = 10). The animals that triggered the collective movement of Amaurobius spiderlings were always larger than the spiderlings. I never observed the young attacking the intruders. On the other hand, no individuals performed contraction when acarids (Acarina) were placed on the web (n = 10).

A single contraction took less than 1 s (0.7–1.0 s). The individual abruptly bent its body while seizing web silk with its legs, and then released it at once. Many individuals within a group performed this action simultaneously. In most cases, the majority of spiderlings within a group participated and their simultaneous movements made the web appear to pulsate strongly.

Synchronous properties of the behavior

Figure 1 presents an example of five contractions performed by ten individuals in the presence of a 20 ± 2 mg cricket nymph. In the first through third contractions shown in Fig. 1, it is clear that the contractions were produced synchronously by the majority of the individuals. The fourth contraction was slightly desynchronized. Successive contractions not interrupted by resting time appeared rhythmic, but the periods of immobility between series of successive contractions were irregular.
https://static-content.springer.com/image/art%3A10.1007%2Fs00040-010-0087-0/MediaObjects/40_2010_87_Fig1_HTML.gif
Fig. 1

An example of contractive movement performed by ten individuals in response to the presence of a 20 ± 2 mg cricket on the web. Data on the graph were obtained from MAS (motion analysis software, Keio Electronic Industrial Co. Ltd.) measurement of the velocity on the video-imaged animal’s movement

The number of individuals participating in contractions varied (Table 1; Chi-square test: df = 6, χ2 = 26.4, P = 0.0002). Not all of the spiderlings participated in every contraction. For example, the second contraction in Table 1 was produced by nine of the ten individuals, while the fourth contraction was produced by a single individual. The frequency of contractions differed among individuals (Chi-square test of independence: df = 9, χ2 = 16.3, P = 0.0600): e.g., individual 8 performed six out of seven successive contractions in the sequence presented in Table 1, while individual 3 did not show any activity in this sequence (Fisher’s exact test: P = 0.0047). Analyses of a number of sequences showed similar results (see below).
Table 1

An observation of 10 spiderlings participating in 7 successive contractions

Individual number

Order of contraction

1st

2nd

3rd

4th

5th

6th

7th

1

C

C

C

C

2

C

C

C

C

3

4

C

C

C

5

C

C

C

C

C

6

C

C

C

C

7

C

C

C

C

C

8

C

C

C

C

C

C

9

C

C

10

C

C

C contraction, – no contraction

The proportion of individuals within a group (n = 20 clutches) participating in a contraction differed among the days post-matriphagy (repeated measures ANOVA test: F3,1208 = 37.67, P < 0.0001): on average, 60.7 (±32.0)% of individuals within a group participated for a contraction on the third day, 46.5 (±30.5)% on the fourth day, 39.2 (±29.6)% on the fifth day and 33.5 (±27.8)% on the sixth day. The proportion of individuals participating was estimated from video sequences of 192 contractions for the third day, 390 images for the fourth day, 284 images for the fifth day and 346 images for the sixth day. There was a progressive decrease in the synchronization of movements among individuals in a group during the period following the matriphagy (Fig. 2; ANOVA for linear regression analysis: F1,1210 = 106.21, y = −8.37x + 82.19, r2 = 0.081, P < 0.0001).
https://static-content.springer.com/image/art%3A10.1007%2Fs00040-010-0087-0/MediaObjects/40_2010_87_Fig2_HTML.gif
Fig. 2

The proportion (%, mean ± SE) of individuals within group participating in a contraction on the days after matriphagy (n = 20 clutches). The mean value was measured for contractions occurring for 10 min following introduction of a 20 ± 2 mg cricket to the web

Ontogeny of the behavior

The phenomenon of contraction appears for only a limited period in development of Am. ferox. Contraction was first observed on the first day post-matriphagy and occurred up to 1 or 2 days after the second molt of the spiderlings (n = 10 clutches). I did not observe contractions in the presence of the mother; the spiderlings showed the behavior for the first time on the first day after matriphagy in response to the introduction of a cricket nymph (20 ± 2 mg) onto the web.

The frequency of contractions produced within 10 min after introduction of the cricket showed clear variation in the course of the days post-matriphagy (Fig. 3). The number of contractions increased until the fourth day and then decreased strongly on the sixth and seventh days. A polynomial parabola fitted to the data showed a maximum on the fourth day post-matriphagy (Fig. 3; ANOVA test for the regression analysis: F2,67 = 13.584, y = −2.58x2 + 18.29x + 5.94, r2 = 0.289, P < 0.0005). There was no difference in the pattern of variation among clutches (ANOVA test: F9,60 = 1.20, P = 0.3200). This result suggests that the appearance and the disappearance of the phenomenon were naturally related to the age of the spiderlings.
https://static-content.springer.com/image/art%3A10.1007%2Fs00040-010-0087-0/MediaObjects/40_2010_87_Fig3_HTML.gif
Fig. 3

Variation over time in the frequency of contractions for 10 min after introduction of a 20 ± 2 mg cricket to the web (n = 10 clutches). The sixth and seventh days corresponded to the period of the spiderlings’ second molt

Factors influencing behavior

External stimuli

An electronic vibrator, used to produce an artificial vibration without an associated chemical stimulus, triggered the contractions of the spiderlings (n = 5 clutches). This experiment also allowed me to assess the effect of variation in the intensity of vibration. The vibration of higher intensity resulted in increased rates of contraction by the spiderlings (Fig. 4; Friedman test: df = 2, χ2 = 6.30, P = 0.0363): the young performed more contractions when the vibration was strong, while they rarely contracted in response to the weak vibration.
https://static-content.springer.com/image/art%3A10.1007%2Fs00040-010-0087-0/MediaObjects/40_2010_87_Fig4_HTML.gif
Fig. 4

Comparison of the number (mean ± SD) of contractions for 3 min as a function of different intensities of vibration produced by an electronic vibrator (n = 5 clutches). The electronic vibrator produced 0.5 vibration per second with a square wave at three different intensities: amplitude of 1.5 cm for ‘strong’ vibrations, 0.7 cm for ‘medium’ and 0.3 cm for ‘weak’

Responses of the individuals appeared to be affected by the movements of the cricket (n = 20 clutches; Table 2). When the cricket moved its whole body, which probably caused strong vibration of the web, the spiderlings showed more contractions. During the absence of movement (even if the cricket was present on the web), I observed minimal contraction. This difference was observed on the third, fourth, fifth, and sixth days post-matriphagy. This result suggests that the rate of production of contractions is related to the intensity of an intruder’s movements on the web.
Table 2

The number of contractions performed by an individual per 100 s according to the movement type of the cricket stimulus (n = 20 clutches)

Day after matriphagy

Movement of the cricket

Friedman test (df = 2)

Body

Appendage

Absent

3rd

5.2 ± 8.2

2.6 ± 6.0

0.1 ± 0.3

χ2 = 8.60, P = 0.0136

4th

8.6 ± 8.5

2.8 ± 3.8

0.1 ± 0.2

χ2 = 8.93, P = 0.0115

5th

3.4 ± 4.1

1.5 ± 1.7

0.1 ± 0.1

χ2 = 9.40, P = 0.0084

6th

2.9 ± 4.0

1.2 ± 2.5

0.1 ± 0.3

χ2 = 15.44, P = 0.0004

Presence of nestmates and the size of the web (the quantity of silk)

The number of contractions produced by an individual, when the cricket was introduced on the web, was significantly higher in the presence of nestmates: I observed 7.0 ± 6.5 contractions in treatment 1, where 10 individuals were present on a web woven by ten individuals (n = 13), and only 2.8 ± 3.4 contractions in treatment 2, where only one individual was present on a web woven by ten individuals (n = 16) (Fig. 5; Mann–Whitney U test: n1 = 13, n2 = 16, U = 166.0, P = 0.0066). On the other hand, one individual on a web woven by ten individuals produced more contractions (2.8 ± 3.4) than one individual on a web woven by one individual (1.3 ± 1.2; n = 15), but this difference was not statistically significant (Mann–Whitney U test: n1 = 16, n2 = 15, U = 154.50, P = 0.1727).
https://static-content.springer.com/image/art%3A10.1007%2Fs00040-010-0087-0/MediaObjects/40_2010_87_Fig5_HTML.gif
Fig. 5

Comparison of the number of contractions per individual in three different treatments. Treatment 1: ten individuals within a web woven by ten individuals (n = 13), treatment 2: one individual within a web woven by ten individuals (n = 16), treatment 3: one individual within a web woven by one individual (n = 15). Data were collected for 3 min following introduction of a 5 ± 0.5 mg cricket to the web on the fifth day after matriphagy

Group density

Individuals in the ‘dense’ group, where the average distance between individuals on the video image was less than the length of an individual, performed significantly more contractions than individuals in the ‘less dense’ groups (Mann–Whitney U test: n1 = 46, n2 = 13, z = 3.2, P = 0.0015). In the dense group (n = 46 clutches), I observed 12.6 ± 17.4 contractions per individual for 10 min following introduction of the cricket, compared with only 1.8 ± 2.4 contractions in the less dense group (n = 13).

Discussion

Am. ferox spiderlings performed the movement of contractions in a synchronized manner among individuals during the post-matriphagy period. Web vibration caused by intruders was the signal releasing the sequence of contractions. Mechanical stimulation from the electronic vibrator also triggered contractions by spiderlings. The intensity of external stimuli affected the number of contractions that the spiderlings produced with more intense vibrations triggering more contractions.

While the strength of the response was related to the intensity of the stimuli, the synchronicity among individuals was not entirely driven by external stimuli. I observed that the spiderlings continued to produce contractions once the behavior started, even when there was no ongoing external stimulus. After contractions are initiated, it is possible that the siblings follow patterns of web movement regardless of the source of movement. This would result in a synchronicity of the individual movements. Additionally, it seemed that the duration of resting time between contractions and the number of contractions were variable within a clutch, despite the fact that in the experimental manipulation, the electronic vibrator provided a stimulus of the same intensity and same frequency throughout the experimental period.

This phenomenon seemed to involve interindividual differences. While contractions were generally performed by multiple individuals, not all of the individuals in a group performed the behavior. The spiderlings might have different thresholds of response to the stimulus. Some individuals might have a stronger tendency to contract than others, which might explain why some individuals respond more than others. Another possibility is that individual responses to the vibration might differ despite similar individual tendencies to respond to the stimulus, because web vibrations could be transmitted to each individual with different intensities (Masters et al., 1986; Craig, 2003). The individual variation may also have been produced by both individual differences and differences in perceived signal intensity.

The results of the analyses of temporal variation in the production of contractions showed that the appearance and disappearance of the phenomenon were related to the age of the spiderlings. The phenomenon appeared from the first day after matriphagy. It continued for a short period of 7–9 days with an increase in frequency up to the fourth day and a decrease on the sixth and seventh days (see Fig. 3). It disappeared a few days after the second molt, when the young moved a short distance in response to an external vibration (personal observation). After the second molt, contraction was performed only by some individuals in the clutch. The synchronicity of the individuals (the number of individuals synchronizing for a contraction) also decreased with the age of the spiderlings (see Fig. 2). The decrease in synchronization with age might result from an increase in the interindividual distance, while distance between siblings on the maternal web increases after matriphagy up to dispersal. This aspect needs to be tested in further investigations.

The mothers are physiologically capable of producing a second clutch, but experiments show that her net reproductive output, calculated as the number of surviving mid-instar juveniles, is maximized by matriphagy versus the alternative strategy of abandoning her progeny early in order to lay a second clutch (Kim et al., 2000). Occurring 1–2 days after the first molt of the young, matriphagy is a stable phenomenon which was observed in all of the clutches investigated in a previous study (Kim and Horel, 1998). The spiderlings did not perform the contractions in the presence of their mother. However, when I manipulated a clutch by removing the mother from the brood just before matriphagy, the young showed collective contraction on the first day after the first molt (pers. obs.). Normally, on the first day after the first molt, spiderlings stay with their mother and do not perform contraction. Matriphagy, therefore, seems not to be necessary for the appearance of the contractive movement. It is possible that the young of second instar are capable of displaying the behavior, but that the behavior is not triggered in presence of the mother.

The spiderlings show predatory activity and generally succeed in capturing a prey item 4–5 days after the second molt (Kim et al., 2005b). Thus, contraction seems to occur only during the period when the mother is not present any more but the young are not yet capable of capturing prey.

The behavior of Am. ferox showed important differences from the successive synchronized stepping toward prey in the non-territorial, permanently social spider, An. eximius (Krafft and Pasquet, 1991; Vakanas and Krafft, 2001). In An. eximius, adult individuals synchronize their movements while advancing toward a prey item, while the movements of Am. ferox spiderlings do not involve directional locomotor activity. Moreover, the behavior of Am. ferox was not followed by any prey-capturing activity. There are also certain similarities. First, the behaviors observed in Am. ferox and An. eximius are collective and synchronized activities. Second, the synchronicity does not involve all of the individuals within group. In addition, these behaviors result from individual responses to an external vibration (mostly provoked by struggling animals).

The adaptive value of the synchronized behavior observed in Am. ferox remains unknown. The phenomenon may be an antipredatory response, such as the collective behaviors of many social animals (e.g., defensive swimming of teleost fishes (Mikheev and Pasternak, 2006), defensive drumming behavior in ant species (Fuchs, 1976), and defensive balling behavior of Apis honeybees (Ichino and Okada, 1994). There are few lines that offer preliminary support for this hypothesis. (1) The spiderlings did not react to a small animal on the web (e.g., the acarids which are probably not a potential predator or possibly not detected by the spiderlings), and they only showed contraction in response to vibrations produced by animals of relatively large size. (2) The spiderlings showed the movements only in the absence of the mother, even though they were able to produce them while she was still with them. (3) The clearly visible movements of the whole web resulting from the behavior of the young might give an intruder the impression that a big organism is nearby. The web movement could also be physically transmitted to the intruder touching the web silk.

Few collective defense behaviors have been observed in spiders. Lubin (1974) recorded a defensive behavior against parasites in the territorial social spider, Cyrtophora moluccencis (Araneidae). When a Hymenoptera flew over a colony of C. moluccencis in search of a cocoon, a female caught her cocoon and shook it strongly. This behavior apparently acted as a signal provoking a similar response from other spiders.

Goss and Deneubourg (1988) developed a model in which autocatalysis was introduced as a possible driver of synchronized and rhythmical activities in social insects. The model was developed to explain the pattern of activities of the ant, Leptothorax acervorum (Formicidae). In this species, the workers have no individual activity rhythm (Franks and Bryant, 1987). Rather, the rate of activation is proportional to the number of active workers, suggesting a positive feedback or autocatalysis. Autocatalysis refers to the idea that the probability of an individual adopting a particular behavior is a function of the number of individuals already exhibiting that behavior (see Krafft and Pasquet, 1991). A decrease in the variability in the intervals between contractions with group size could provide indirect evidence that the synchronization arises through interactions among spiderlings. This needs further studies.

The fact that Am. ferox presents the contraction alone shows that the behavior is not a phenomenon programmed in the group level and that feedback from the group is not compulsory in an individual contraction. Nevertheless, the positive effects of the nestmates and high group density (i.e., short distance between individuals) suggest that the presence of siblings tends to amplify individual contractions.

Acknowledgments

I am deeply grateful to André Horel, Chantal Roland, Valérie Grasmück, Samuel Venner and Laurent Thévenard for their help while I conducted the experiments and for helpful discussions. I thank also Hyun Shin, Susan Lappan and Raphaël Jeanson for invaluable comments on the manuscript. This work was partially supported by the University of Incheon, South Korea. The study was conducted in compliance with ethical standards of animal treatment according to the Association for the Study of Animal Behaviour/Animal Behavior Society Guidelines for the use of animals in research.

Copyright information

© International Union for the Study of Social Insects (IUSSI) 2010