- Email: [email protected]
Prey orientation and the role of venom availability in the predatory behaviour of the centipede Scolopendra subspinipes mutilans (Arthropoda: Chilopoda)
Journal of Insect Physiology 58 (2012) 874–880
Contents lists available at SciVerse ScienceDirect
Journal of Insect Physiology journal homepage: www.elsevier.com/locate/jinsphys
Prey orientation and the role of venom availability in the predatory behaviour of the centipede Scolopendra subspinipes mutilans (Arthropoda: Chilopoda) Michel M. Dugon ⇑, Wallace Arthur Department of Zoology, School of Natural Sciences, National University of Ireland, University Rd, Galway, Ireland
a r t i c l e
i n f o
Article history: Received 8 February 2012 Received in revised form 23 March 2012 Accepted 26 March 2012 Available online 7 April 2012 Keywords: Centipede Venom Prey selection Venom optimization hypothesis Predatory behavior
a b s t r a c t Many animal phyla contain clades in which most or all species are venom-injecting predators. An example, in the arthropods, is the class Chilopoda, containing the approximately 3500 species of centipedes. Very little ecological or behavioural work yielding quantitative data has been conducted on centipede predation. Here, we describe a study of this kind. Our experiments employed one centipede species – a large tropical one, Scolopendra subspinipes mutilans – and two species of prey – a cricket, Gryllus assimilis, and a locust, Schistocerca gregaria. We conducted two experiments. The first was aimed at investigating the extent to which the centipedes attacked prey in particular tagmata as opposed to at random over the whole body surface. The results showed that the centipedes were highly selective, preferring to attack the head or thorax rather than the abdomen; indeed, they often reoriented the prey in order to achieve this. A possible explanation of this behaviour is to maximize the speed with which the neurotoxins in the venom reach either the brain or the thoracic ganglia that control limb movement. The second experiment was aimed at investigating the effect of venom-extraction on the attack rate, and specifically at testing if the magnitude of any such effect differed between the two types of prey, which differ considerably in size. The results showed a major effect of venom extraction in relation to both types of prey, but with the time taken to return to a ‘normal’ attack rate being longer in the case of the larger prey-type, namely the locust. We discuss these results in relation to the ‘venom optimization hypothesis’ and, more generally, to the principle of minimizing the production/use of venom, which is an energetically expensive resource. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The class Chilopoda (Arthropoda: Myriapoda) contains approximately 3500 described species of centipedes distributed among five living orders (Edgecombe and Giribet, 2007). Centipedes are predators characterised by a dorsoventrally flattened body bearing 15 or more pairs of legs, one pair per trunk segment. The head is attached to the trunk, which is not divided into thorax and abdomen. The first post-cephalic segment bears a pair of strong and sharp claws, named forcipules, each containing a venom system. Proximally, the venom-producing gland is encapsulated deep into the forcipule. Distally, the forcipule finishes into a sharp apical claw bearing a subterminal opening (the meatus) through which the venom is released (Fig. 1) (Dugon and Arthur, 2012). With the exception of a few species (e.g.: the scolopendromorph Paracryptops spinosus with unusually short forcipules – see Jangi and Dass, 1978), centipedes are thought to feed predominantly on live animals (Voigtländer, 2011). Centipedes seize their prey with the anterior legs and usually envenomate them with a stabbing ⇑ Corresponding author. E-mail address: [email protected] (M.M. Dugon). 0022-1910/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jinsphys.2012.03.014
motion of the forcipules. Plant material is also sometimes ingested but this intake is thought to be negligible (Lewis, 1981). Centipedes are considered opportunistic predators, catching any prey that can be overpowered. Large centipedes of the order Scolopendromorpha, the ‘giants’ of the centipede world, reaching up to 30 cm long, do sometimes attack large arthropods (scorpions; Iorio, 2004), amphibians (toads: Carpenter and Gillingham, 1984), reptiles (small snakes: Okeden, 1903 reported by Lewis, 1981), and also small mammals (mice: Voigtländer, 2011; bats: Molinari et al., 2005). Little is known about the techniques used by centipedes to capture prey items. Scolopendromorphs display a variety of hunting methods, from wandering on the leaf-litter, flicking the anterior half of the body on both sides in order to make contact with the prey and outrun it, to actively catching prey in mid-flight, sometimes hanging upside down from the ceiling of a cave (Molinari et al., 2005). The scolopendrid Scolopendra viridis has been reported to consistently use the anterior walking legs to hold the prey while venom is delivered with the forcipules (Elzinga, 1994). Several species have been observed seizing prey with the terminal pair of legs, and then flicking the anterior half of the body to inject the prey with venom.
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
875
Fig. 1. Left: Scolopendra sp. from Hubei, China, preening the antennae using the 2nd pair of maxillae. The forcipules are clearly visible as a pair of sharp claws curving anteromedially under the head capsule. Right: Scolopendra subspinipes mutilans attacking a cricket. The forcipules, having stabbed the prey, are holding it firmly in place while the venom takes effect.
Venom is a key element in the predatory behaviour of centipedes. Whether venom is solely used for immobilising prey by inducing paralysis remains unclear. It has been proposed that venom also contains digestive enzymes used to soften-up the flesh of the prey which is subsequently sucked-up (Undheim and King, 2011). The exoskeleton of the prey is usually not consumed. So far, venomic analyses have been performed only on large scolopendromorphs known to have a medically significant bite. Noda et al. (2001) isolated the benzoic alkaloid Scolopendrine as well as serotonic and muscarinic components, the latter two being powerful pain triggers and toxic catalysers. The neurotoxicity of Scolopendra sp. venom has been demonstrated by Gutierrez et al. (2003). Rates et al. (2007) and Malta et al. (2008) provided the most recent comparative study on centipede venom, based on South-American scolopendromorphs. The authors showed the myotoxic and weak haemolytic properties of the venoms. Their findings corroborate the observations of Bücherl (1946, 1971) on small mammals injected with diluted venom. Arthropods injected with venom are disabled within seconds (Voigtländer, 2011), a result expected in the presence of neurotoxins. The presence of myotoxic and haemolytic effects may be congruent with the hypothesis of external digestion. Undheim and King (2011) recently reviewed the literature pertaining to the composition and toxicity of centipede venoms. The behavioural processes involved in prey capture and venom injection in centipedes are virtually unknown, even in the largest species of scolopendromorphs. Physiological experiments show that stimulation of the forcipular sensilla may trigger a mechanical feeding response (Jangi and Dass, 1977). However, nothing is known about the role of venom availability in the general feeding response of centipedes. Here we investigate two aspects of predatory behaviour in the Chinese red-headed centipede Scolopendra subspinipes mutilans (Scolopendromorpha: Scolopendridae). The first experiment consisted of recording the behaviour of prey capture and evenomation. A second experiment was designed to investigate the relationship between predatory behaviour, venom availability and prey-type. 2. Materials and methods Two distinct experiments were performed: 1. A prey orientation experiment, in which 50 centipedes were presented each in turn with a prey item (the field cricket Gryllus assimilis). The reaction of each centipede was recorded and later analysed in order to assess if S. s. mutilans reorientate the prey after the initial seizure to inject venom in a ‘favoured’ part of the prey.
2. An experiment investigating prey attack in relation to venom availability. 150 S. s. mutilans were used in this experiment, in which the centipedes were offered one or another of two types of prey items ( G. assimilis or the migratory desert locust Schistocerca gregaria) at different times after venom extraction. 3. General set-up Prior to experiments 1 and 2, the centipedes were starved for 2 weeks. As centipedes are thought to be mainly nocturnal predators, each set of treatments (for both experiments 1 and 2) started at about 10 pm. The experimental ‘arena’ consisted of a translucent acrylic container (200 mm 140 mm 115 mm). No substrate was used and the container was briefly cleaned before introducing a new specimen into the arena. No artificial light was directed at the arena, and the observations were performed using dimmed, ceiling-fixed, fluorescent tubes. The experiments were performed using one centipede and one prey item at a time. Each centipede was introduced into the arena and was allowed to settle for 2 min. A prey item was then introduced into the enclosure containing the centipede. If the prey and the centipede were immobile, the prey was induced, with a gentle brush stroke, to walk towards the centipede. The behaviour of the centipede was recorded for 3 min. At the end of the 3 min, the centipede was removed from the experimental arena and returned to its normal container. 4. Prey items The field crickets and the desert locusts were purchased online from Livefood, UK. Large specimens of both prey species were selected for the experiments. From the base of the antennae to the end of the abdomen, the selected G. assimilis measured a minimum of 18 mm and a maximum of 22 mm (average: 20 mm). This resulted in a weight range between 400 and 500 mg (average: 438 mg). The selected specimens of S. gregaria measured at least 40 mm (average: 48 mm) and weighed between 1300 and 2300 mg (average: 1561 mg). Both prey items belong to the order Orthoptera and are considered to be similarly vulnerable to the venom of Scolopendra species (Iorio, 2004). However, as S. gregaria is much larger and heavier than G. assimilis, it was assumed that more venom would be necessary to tackle a S. gregaria. 4.1. Experiment 1: prey orientation experiment The goal of this experiment was to determine if centipedes manipulate their prey (in this case always a G. assimilis) in order to orientate them in a particular fashion prior to venom delivery. Consistency in the orientation of the prey and consistency in the
876
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
tagma into which venom is delivered may indicate that centipedes inject venom where it is most effective. 50 specimens of S. s. mutilans were tested in this experiment. Up to four behavioural events were recorded: Prey attacked/not attacked: A successful attack consisted of the prey being seized with the forcipules and not released during the 3 min of observation, regardless of subsequent manipulations. Prey seized and then released: Unsuccessful attacks occurred when the prey was seized and subsequently released within the 3 min of observation. Location of the initial bite: This consists of recording the part of the body of the prey into which the forcipules were first inserted during the seizure. Two body parts are distinguished on the prey: (1) the head + thorax; (2) the abdomen. Location of the final bite: This is the location of the forcipules at the end of the 3 min of observation. If manipulations occurred, the location of the final bite may be different from the initial bite. Only movements involving a change of bite location from head + thorax to abdomen or from abdomen to head + thorax were recorded as a changed location. If centipedes insert their forcipules anywhere indiscriminately on the prey, the expected frequencies for bite location (head + thorax or abdomen) should not be significantly different from the proportions these parts occupy on the prey’s body. These proportions were assessed by measuring the length of the head + thorax and the length of the abdomen in ten specimens, both on their dorsal and ventral sides. The values were then averaged. In the case of G. assimilis, the head + thorax represented about 37% of the whole body length, the abdomen representing the remaining 63%. Therefore, in the case of spatially random envenomation, 37% of bites were expected to be delivered on the anterior tagmata and 63% were expected to occur on the abdomen. The significance of the results was tested using either a chisquare goodness-of-fit test (Table 3) or a chi-square test for independence (Table 4). The analyses were performed using Graphpad (http://www.graphpad.com).
4.2. Experiment 2: venom availability and prey attack rate experiment This experiment was designed to test if the attack rate on two types of prey items differs significantly depending on the amount of venom available in the venom glands of the centipede. The experiment consisted of extracting venom and then giving either a small prey (the cricket) or a larger prey (the locust) to different centipedes at different times after venom extraction. If centipedes have the ability to ‘decide’ whether to attack prey depending on the venom available in the venom gland, crickets should be attacked sooner after venom extraction than locusts. In this experiment, 150 centipedes were divided into six groups of 25 specimens each, as listed below. Four groups of centipedes had their venom extracted, and two groups were used as negative control groups, one for each type of prey. Prior to venom extraction, the centipedes were anaesthetised by CO2 exposure. Then, each centipede was attached, ventral side facing up, to an acrylic sheet using two pieces of adhesive tape. Venom was extracted by electro-stimulation using a double pulse stimulator (Cat. 6032, Scientific and Research Instruments Ltd.). Each electrical impulse lasted for 5 ms at 22 V and recurred every 110 ms at the same voltage. The tips of the electrical probes were equipped with wet sponges regularly immersed in a saline solution. The sponges were applied to the sides of the forcipules for approximately 2 s. This operation was repeated 3–4 times in order to extract as much venom as possible. The venom was collected using a pipette between each repeat in order to clean the area and to confirm good extraction of the venom at each application
of the probes. The specimen was then removed from the sheet and replaced into its enclosure. Control groups assured that the feeding response was not affected by the anaesthesia and electric shocks. The specimens were treated as for a venom extraction but electric impulses were never applied on the forcipules. Instead, one electrical probe was applied on the third leg-bearing segment and the other on the twentieth segment so that although the current ran through most of the body, it did not pass through the anteriormost segments or the head capsule. Any application of the probes on the first or second leg-bearing segment would result in venom being discharged. The six groups of 25 centipedes each were treated as follows (see also Table 1): Group 1 (control): G. assimilis were given 6 h after anaesthesia and electrocution. Group 2: G. assimilis were given 6 h after anaesthesia and venom extraction. Group 3: G. assimilis were given 24 h after anaesthesia and venom extraction. Group 4 (control): S. gregaria were given 6 h after anaesthesia and electrocution. Group 5: S. gregaria were given 24 h after anaesthesia and venom extraction. Group 6: S. gregaria were given 48 h after anaesthesia and venom extraction. The prey-seizure frequencies were recorded for each group. The following between-group prey-seizing frequencies were compared in 2 2 contingency tables using a chi-square test with one degree of freedom: Group 1 (neg. ctrl/6 h/G. assimilis) versus group 2 (ven. ext./6 h/ G. assimilis). Group 2 (ven. ext./6 h/G. assimilis) versus group 3 (ven. ext./ 24 h/G. assimilis). Group 1 (neg. ctrl/6 h/G. assimilis) versus group 4 (neg. ctrl/6 h/ S. gregaria). Group 3 (ven. ext./24 h/G. assimilis) versus group 5 (ven. ext./ 24 h/S. gregaria). Group 4 (neg. ctrl/6 h/S. gregaria) versus group 5 (ven. ext./24 h/ S. gregaria). Group 5 (ven. ext./24 h/S. gregaria) versus group 6 (ven. ext./ 48 h/S. gregaria).
The results of experiment 2 were analysed using both chisquared and Fisher’s exact tests.
5. Results 5.1. Experiment 1: prey orientation prior to envenomation Of a total of 50 S. s. mutilans that were offered an adult G. assimilis, 49 (98%) seized the prey within the 3 min observation timeframe. A single centipede (2%) showed no interest in the cricket. None of the centipedes subsequently released the prey within the 3 min. Once seized, the prey was reoriented in 14 cases (29%) (Table 2). 28 (57%) of the 49 seizure events occurred on the head + thorax. The abdomen was seized in 21 cases (43%). As can be seen from table 3, these observed frequencies are significantly different from the expected ones, indicating that centipedes preferentially attack the head and thorax rather than the abdomen during the initial seizure.
877
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
Table 1 Treatment details for 6 groups of adult Scolopendra subspinipes mutilans in the venom availability experiment. Each group is composed of 25 specimens starved for 2 weeks prior to the start of the experiment. Group 1 Prey item Treatment Resting period before feeding
Group 2
Gryllus assimilis (cricket) Control Venom extraction 6h 6h
Group 3
Group 4
Venom extraction 24 h
Schistocerca gregaria (locust) Control Venom extraction 6h 24 h
Prey not attacked 1 Prey not released 49 Prey not reoriented 35
Group 6 Venom extraction 48 h
reorientate their prey in order to inject venom into the head + thorax rather than into the abdomen.
Table 2 Attack frequency of Scolopendra subspinipes mutilans on Gryllus assimilis. Prey attacked 49 Prey released 0 Prey reoriented 14
Group 5
Total 50 Total 49 Total 49
5.2. Experiment 2: prey attack rate and venom availability
The prey was not reoriented in 71% of successful attacks. Among these, 77% of bites occurred on the head + thorax of G. assimilis; the abdomen was seized in the remaining 23% of cases (Table 3). This demonstrates the preference of S. s. mutilans for biting the anterior tagmata of the prey. Reorientation occurred from head + thorax to the abdomen in 7% of all reorientations, and from the abdomen to the head + thorax in the remaining 93%. The expected frequencies of bite location (on the basis of random manipulation) were determined by the relative sizes of the head + thorax and abdomen in G. assimilis (Table 3). These expected frequencies were not confirmed, thus inferring that reorientation of the prey is significantly biased towards a secondary bite located on the head + thorax. At the end of the 3 min of observation, and regardless of reorientation, 82% of the final bites were located on the head + thorax and 18% were located on the abdomen. The high significance of the departure from the expected bite location frequencies (Table 3) demonstrates that centipedes do not bite the prey randomly but definitely favour venom injection in the head or thorax. The difference between the observed frequencies of initial and final bite locations was highly significant (Table 4). This is consistent with previous results and with the hypothesis that centipedes
5.2.1. G. assimilis as a prey item When venom was not extracted, 96% the centipedes attacked the prey. A single specimen (4%) disregarded the cricket. None of the centipedes that attacked subsequently released the prey. When the observed attack frequencies of the G. assimilis control group are compared to the observed attack frequencies in the prey orientation experiment, it can be seen that the anaesthesia and the electric impulses do not significantly change the predatory behaviour of the centipedes (Table 5A). Six hours after venom extraction, only 16% of the centipedes bit the prey-cricket and held onto it. Twenty-four percent of centipedes seized and then released the prey. The prey was totally ignored in the remaining 40% of cases. The ‘prey-seized-andreleased’ cases are treated as ‘prey not attacked’. When comparing group 1 (G. assimilis – Control) versus group 2 (G. assimilis – 6 h extraction), it appears that the frequency of successful attacks 6 h after venom extraction is significantly lower than the frequency of successful attacks when venom is not extracted (Table 5B). In the event of a ‘prey-seized-and-released’, most released prey crickets are capable of fleeing despite the stabbing wounds caused by the sharp apical claws of the centipede legs involved in seizing. This observation is consistent with the expected non-effectiveness of the venom apparatus 6 h after venom extraction. 24 h after venom extraction, 92% of centipedes belonging to the group 3 (G. assimilis – 24 h extraction) seized and held onto the prey. One specimen performed a ‘seize-and-release’ and another one ignored the prey completely. These results are significantly
Table 3 Expected and observed frequencies of initial seizure and envenomination of a prey cricket Gryllus assimilis by Scolopendra subspinipes mutilans. Out of 50 trials, initial seizure of the prey by the centipede occurred 49 times. In 35 cases, the prey was subsequently envenomated without any further manipulation. Reorientation of the prey occurred in 14 cases. The final envenomation indicates the location of the definitive bite at the end of the 3 min of observation. Expected frequencies were calculated on the basis of random seizure or manipulation, proportionally to the relative size of the tagmata in relation to whole body length.
Initial seizure (N = 49) Envenomation location, no reorientation (N = 35) Reorientation + envenomation (N = 14) Final envenomation location (N = 49)
Location
Expected frequencies
Observed frequencies
X2 Goodness-of-fit
Head + thorax Abdomen Head + thorax Abdomen From head + thorax to abdomen From abdomen to head + thorax Head + thorax Abdomen
18 31 13 22 9 5 18 31
28 21 27 8 1 13 40 9
X2 = 8.781 p = 0.003 X2 = 23.986 p < 0.0001 X2 = 19.119 p < 0.0001 X2 = 42.502 p < 0.0001
Table 4 Observed frequencies of the initial and final bite locations following successful attacks on Gryllus assimilis by Scolopendra subspinipes mutilans.
Initial bites Final bites
Bites located on the head + thorax
Bites located on the abdomen
Total
Chi square for independence test
28 40
21 9
49 49
X2 = 6.918 p < 0.01
878
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
Table 5 Observed attack frequencies of Scolopendra subspinipes mutilans on two types of prey items at different times after venom extraction. Group 1: control. Gryllus assimilis given 6 h after anaesthesia and electric shocks. Group 2: Gryllus assimilis given 6 h after venom extraction. Group 3: Gryllus assimilis given 24 h after venom extraction. Group 4: control. Schistocerca gregaria given 6 h after anaesthesia and electric shocks. Group 5: Schistocerca gregaria given 24 h after venom extraction. Group 6: Schistocerca given 48 h after venom extraction. Prey attacked
Prey not attacked
Total
X2
P
1
50
0.26
ns
32.47
p < 0.0001
29.07
p < 0.0001
35.28
p < 0.0001
13.23
p < 0.001
0.36
ns
35.28
p < 0.0001
A
Experiment 1 Vs. Group 1 (G.assimilis – Control)
49 24
1
25
B
Group 1 (G. assimilis – Control) Vs. Group 2 (G. assimilis – 6 h extraction)
24
1
25
4
21
25
C
Group 2 (G. assimilis – 6 h extraction) Vs. Group 3(G. assimilis – 24 h extraction)
4
21
25
23
2
25
D
Group 4 (S. gregaria – Control) Vs. Group 5 (S. gregaria – 6 h extraction)
23
2
25
2
23
25
E
Group 5 (S. gregaria – 24 h extraction) Vs. Group 6(S. gregaria – 48 h extraction)
2
23
25
14
11
25
F
Group 4 (S. gregaria – Control) Vs. Group 1 (G.assimilis – Control)
23
2
25
24
1
25
G
Group 3 (G. assimilis – 24 h extraction) Vs. Group 5 (S. gregaria – 24 h extraction)
23
2
25
2
23
25
different from the results obtained 6 h after venom extraction (group 2), as shown in Table 5C. The results are very similar to those obtained with the control group, when venom was not extracted.
6. Discussion
5.2.2. Schistocerca gregaria as a prey item When no venom was extracted (group 4), 92% of centipedes attacked the locust 6 h after electrocution and 8% showed no interest in it. None of the centipedes that made an attack subsequently released the prey. Twenty-four hours after venom extraction, only 8% of centipedes seized and held their prey. In 36% of cases, the prey was seized and then released. Interestingly, the ‘seized-and-released’ locusts appeared (at least superficially) unharmed, and were capable of retaliation by ‘kicking’ the centipedes with the posterior legs. Fifty-six percent of centipedes ignored the prey. If the ‘prey-seizedand-released’ cases are treated as ‘prey not attacked’ and the results of this (group 5) are tested against the results of the control (group 4), it can be seen that the frequency of successful attacks 24 h after venom extraction is significantly lower than when venom is not extracted (Table 5D). Forty-eight hours after venom extraction (group 6), 56% of centipedes seized and held onto their prey; 24% performed a ‘seizeand-release’. Twenty percent of centipedes did not attack the prey at all. The frequency of successful attacks significantly increased at 48 h compared to the frequency at 24 h after venom extraction (Table 5E). However, the number of successful attacks was not at the level seen in the control group in which venom was not extracted.
When seizing a prey cricket, a centipede usually bites its head and thorax rather than its abdomen. A bite located on the abdomen is usually followed by a reorientation of the prey and by a secondary bite on the head or thorax. The number of final bites on the anterior tagmata is significantly higher than the expected number. This observation leads to two questions. First, why would centipedes show a preference for injecting venom in the anterior end of the prey rather than the posterior? And second, how do centipedes spatially comprehend both their own position and the position of the prey? The main role of venom is to disable prey in an effective and prompt manner (the second role being the lysis of internal tissues). The venoms of Scolopendridae are known to contain, among many other elements, neurotoxins disrupting synaptic signals (Stankiewicz et al., 1999; Gutierrez et al., 2003). We can suspect that these neurotoxins are the major fast-acting disabling elements of many centipedes’ venoms. By changing the properties of the synaptic channels, neurotoxins may act as strong paralytics when reaching the central nervous system. Therefore, the venom is most efficient when injected either into the head capsule or close to the thoracic ganglia controlling locomotion. In the case of insects these ganglia are present on the ventral side of the thorax. When injected into the soft tissues of the abdomen, it is likely that the venom would be diluted in the circulatory system and the gut of the prey, and would have to travel further in order to disable locomotion and prevent the escape of or retaliation by the prey. However, considering the amount and the potency of the venom delivered by large scolopendromorphs (see Bücherl, 1946, 1971), it is very likely that the location of venom injection does not matter much in order to disable a small or average-sized prey item. Prey orientation may be based on two complementary principles. The first is a principle of economy: venom is physiologically expensive to produce and the replenishment of the venom glands requires time. The second principle is based on limiting prey retaliation: seizing the prey with the walking legs in the alignment of
5.2.3. Between prey-type analysis When the results of the control group with G. assimilis preytype (group 1) are tested against the control group with S. gregaria prey-type (group 4), they are not significantly different (Table 5F). The anaesthesia and the electric impulses do not significantly change the predatory behaviour of the centipedes, and S. s. mutilans attacks both prey types at the same rate. A comparison of the attack frequencies between centipede groups that were given G. assimilis (group 3) or S. gregaria (group 5) 24 h after venom extraction showed that centipedes attack the crickets more than the locusts (Table 5G).
6.1. Prey orientation experiment
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
the anterior-posterior body axis while stabbing its thorax or head with the forcipules restrains the prey sufficiently to avoid retaliation until the venom takes effect. The results of the prey orientation experiment also provide some information on the capacity of centipedes to perceive the three-dimensionality of their surroundings. Because the primary attack occurs within a few tenths of a second, it is unlikely that the centipede always manages to ‘understand’ the spatial orientation of the prey through physical contact with the antennae or another body part prior to the attack. Spatial orientation must also occur through other means. First, it is probable that the perception of the prey’s orientation takes place once the centipede has pounced on the prey and ‘felt’ the prey with its legs and immobilised it, prior to any forcipular movement. This would explain the occurrence of pre-bite manipulations and the systematic use of the walking legs during the grasping motion. Second, we could consider that, to a limited extent, S. s. mutilans uses its eyesight at very short distances to perceive a prey’s orientation before the attack. However, a specific role of eyesight in prey capture by scolopendromorphs is unlikely in view of previous reports on the subject, which all suggest contact, mechanoreception or chemoreception as the main tools for prey recognition and localisation (Cloudsley-Thompson, 1955 mentioned by Lewis, 1981; Manton, 1965). 6.2. Venom availability and prey attack rate experiment Overall, the results observed in experiment 2 demonstrate two distinct patterns: First, it appears that the hunting behaviour of centipedes is linked to venom available in the glands. The behavioural pattern is altered in two ways by reduced venom: (1) a large proportion of centipedes refuse to attack the prey; and (2) some centipedes release their prey after seizing it, a phenomenon not observed when there was no venom extraction. The physiological mechanisms allowing centipedes to perceive the amount of venom available and subsequently to change their behaviour is still unknown. However, the occurrence of failed attacks suggests also that centipedes may sometimes ‘override’ the physiological feedback and engage themselves in an attempt to attack a prey. Because the centipedes used in these experiments were starved for 2 weeks, it is possible that hunger prevailed over lack of venom availability and the risks linked to retaliation by the prey. Also, it should be noted that centipedes involved in failed attacks released their prey after a few seconds. It is possible that the mechanoreceptors and chemoreceptors located on the tarsungulum (distal end) of the forcipules provide feedback on the amount of venom discharged and the condition of the prey in the seconds following the bite. In Scolopendra morsitans, the application of a glucose solution on the sensilla of the tarsungulum triggers a response of the feeding apparatus, presumably acting via chemoreceptors (Jangi and Dass, 1977). The second distinct pattern is the time lapse between venom extraction and the return towards a ‘normal’ hunting response. When the prey is small (G. assimilis, with an average weight of 438 mg), it seems that centipedes have produced enough venom to tackle the prey normally 24 h after venom extraction. When the prey is much larger (S. gregaria, with an average weight of 1561 mg), most centipedes refuse to attack the prey 24 h after venom extraction. It is only 48 h after venom extraction that the frequency of successful attacks increases significantly and even then it has not returned to its ‘normal’ level. These observations demonstrate that centipedes do perceive the difference between small and large prey, and alter their behaviour accordingly. The amount of venom available, the size of the prey and possibly the resistance of the prey to the venom may all be taken into account before any ‘decision’ is made about attacking. But once again the number of
879
failed attacks tempers this idea, as wrong ‘decisions’ lead sometimes to unsuccessful attempts, a phenomenon not observed when the venom gland is full. Venom-economy strategies by means of selective venom injection and prey orientation have been previously observed and investigated only in very few species of terrestrial arthropods, such as the wandering spider Cupiennius salei (Wullschleger and Nentwig, 2002; Wigger et al., 2002; Kuhn-Nentwig et al., 2004; Hostettler and Nentwig, 2006; Boevé, 1994), the New-Zealand crab spider Diaea sp. (Pollard, 1990), the scorpion Parabuthus transvaalicus (Nisani and Hayes, 2011), the arboreal African ponerine ant Platythyrea conradti (Dejean, 2011) and a predatory wasp (Haspel et al., 2003). Studies have shown that P. conradti turns its prey over and stings them almost exclusively on the ventral part of the thorax, where the neural cord is closest to the exoskeleton. Also, P. conradti injects paralytic venom only into large prey presenting a retaliation potential (Dejean, 2011). Some predatory wasps inject venom specifically into the brain of their prey (Haspel et al., 2003). This principle of venom economy has been named the venom optimization hypothesis (Wigger et al., 2002). In the case of S. s. mutilans (and P. conradti for that matter), the venom optimization occurs in two ways: (1) orientate the prey in order to deliver venom in the most sensitive area; and (2) attack a prey matching the amount of venom available in the gland. Interestingly, the few species in which venom optimisation has been observed belong to three different subphyla of Arthropoda (Chelicerata, Hexapoda and now Myriapoda). Given the diverse scattering of major venomous groups (like centipedes) across the arthropod evolutionary tree, it seems likely that venom optimization has evolved convergently several times. In each case, the basis of the evolutionary trend may be the principle of economising venom, an expensive resource. Acknowledgements We thank Eoin MacLoughlin and ‘Uncle’ Daniel Leong for logistic support. This work was funded by a Professor Brendan F. Keegan memorial Ph.D. studentship (held by M.D.) from Eli Lilly, to whom we are grateful for their financial support. References Boevé, J.L., 1994. Injection of venom into an insect prey by the free hunting spider Cupiennius salei (Araneae, Ctenidae). Journal of Zoology 234, 165–175. Bücherl, W., 1946. Açâo do veneno dos escolopendromorfos do Brasil sobre alguns animais de laboratorio. Memórias do Instituto Butantan 19, 181–197. Bücherl, W., 1971. Venomous chilopods or centipedes. In: Bücherl, W. and Buckley, E.E. (Eds.), Venomous Animals and Their Venoms. Academic Press, New-York, pp. 169–196. Carpenter, C.C., Gillingham, J.C., 1984. Giant centipede (Scolopendra alternans) attacks marine toad (Bufo marinus). Caribbean Journal of Sciences 20, 71–72. Cloudsley-Thompson, J.L., 1955. Some aspects of the biology of centipedes and scorpions. The Naturalist 6, 147–153. Dejean, A., 2011. Prey capture behavior in an arboreal African ponerine ant. Public Library of Science One 6, e19837. Dugon, M.M., Arthur, A., 2012. Comparative studies on the structure and development of the venom-delivery system of centipedes; and a hypothesis on the origin of this evolutionary novelty. Evolution & Development 14, 1– 10. Edgecombe, G.D., Giribet, G., 2007. Evolutionary biology of centipedes (Myriapoda: Chilopoda). Annual Review of Entomology 52, 151–170. Elzinga, R.J., 1994. The use of legs as grasping structures during prey capture and feeding by the centipede Scolopendra viridis Say (Chilopoda: Scolopendridae). Journal of the Kansas Entomological Society 67, 369–372. Gutierrez, M.C., Abarca, C., Possani, L.D., 2003. A toxic fraction from Scolopendra venom increases the basal release of neurotransmitters in the ventral ganglia of crustaceans. Comparative Biochemistry and Physiology 135, 205–214. Haspel, G., Rosenberg, L.A., Libersat, F., 2003. Direct injection of venom by a predatory wasp into cockroach brain. Journal of Neurobiology 56, 287–292. Hostettler, S., Nentwig, W., 2006. Olfactory information saves venom during preycapture of the hunting spider Cupiennius salei (Araneae: Ctenidae). Functional Ecology 20, 369–375.
880
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
Iorio, E., 2004. Les appareils venimeux des Chilopodes: mécanismes et pathologies. Bulletin de Phyllie 20, 23–33. Jangi, B.S., Dass, C.M.S., 1977. Chemoreceptive function of the poison fang in the centipede Scolopendra morsitans Linn. Indian Journal of Experimental Biology 15, 803–804. Jangi, B.S., Dass, C.M.S., 1978. A new species of the genus Paracryptops from India, with remarks on the generic and specific characters (Chilopoda: Scolopendromorpha: Cryptopidae). Zoological Journal of the Linnean Society 64, 327–330. Kuhn-Nentwig, L., Schaller, J., Nentwig, W., 2004. Biochemistry, toxicology and ecology of the venom of the spider Cupiennius salei (Ctenidae). Toxicon 43, 543– 553. Lewis, J.G.E., 1981. The Biology of Centipedes. Cambridge University Press, UK. Malta, M.B., Lira, M.S., Soares, S.L., Rocha, G.C., Knysak, I., Martins, R., Guizze, S.P.G., Santro, M.L., Barbaro, K.C., 2008. Toxic activities of Brazilian centipede venoms. Toxicon 52, 255–263. Manton, S.M., 1965. The evolution of arthropodan locomotory mechanisms. Part 8. Functional requirements and body design in Chilopoda. Journal of the Linnean Society of London 46, 251–484. Molinari, J., Gutierrez, E.E., De Ascencao, J.M.N., Arends, A., Marquez, R.J., 2005. Predation by giant centipedes, Scolopendra gigantea, on three species of bats in a Venezuelan cave. Caribbean Journal of Sciences 41, 340–346. Nisani, Z., Hayes, W.K., 2011. Defensive stinging by Parabuthus transvaalicus scorpions: risk assessment and venom metering. Animal Behaviour 81, 627– 633.
Noda, N., Yashiki, Y., Nakatani, T., Miyahara, K., Du, X.-M., 2001. A novel quinoline alkaloid possessing a 7-benzyl group from the centipede, Scolopendra subspinipes. Chemical & Pharmaceutical Bulletin 49, 930–931. Okeden, W.P., 1903. A centipede eating a snake. Journal of Bombay Natural History Society 15, 1. Pollard, S.D., 1990. The feeding strategy of a crab spider, Diaea sp. indet. (Araneae: Thoisidae): post-capture decision rules. Journal of Zoology 222, 601–615. Rates, B., Bemquerer, M.P., Richardson, M., Borges, M.H., Morales, R.A.V., De Lima, M.E., Pimenta, A.M.C., 2007. Venomic analyses of Scolopendra viridicornis nigra and Scolopendra angulata (Centipede, Scolopendromorpha): shedding light on venoms from a neglected group. Toxicon 49, 810–826. Stankiewicz, M., Hamon, A., Benkhalifa, R., Kadziela, W., Hue, B., Lucas, S., Mebs, D., Pelhate, M., 1999. Effects of a centipede venom fraction on insect nervous system, a native Xenopus oocyte receptor and on an expressed Drosophila muscarinic receptor. Toxicon 37, 1431–1445. Undheim, E.A.B., King, G.F., 2011. On the venom system of centipedes (Chilopoda), a neglected group of venomous animals. Toxicon 57, 512–524. Voigtländer, K., 2011. Chilopoda – Ecology. In: Minelli, A. (Ed.), The Myriapoda, vol. 1. Brill, Leiden, pp. 309–325. Wigger, E., Kuhn-Nentwig, L., Nentwig, W., 2002. The venom optimisation hypothesis: a spider injects large venom quantities only into difficult prey types. Toxicon 40, 749–752. Wullschleger, B., Nentwig, W., 2002. Influence of venom availability on a spider’s prey-choice behaviour. Functional Ecology 16, 802–807.
Contents lists available at SciVerse ScienceDirect
Journal of Insect Physiology journal homepage: www.elsevier.com/locate/jinsphys
Prey orientation and the role of venom availability in the predatory behaviour of the centipede Scolopendra subspinipes mutilans (Arthropoda: Chilopoda) Michel M. Dugon ⇑, Wallace Arthur Department of Zoology, School of Natural Sciences, National University of Ireland, University Rd, Galway, Ireland
a r t i c l e
i n f o
Article history: Received 8 February 2012 Received in revised form 23 March 2012 Accepted 26 March 2012 Available online 7 April 2012 Keywords: Centipede Venom Prey selection Venom optimization hypothesis Predatory behavior
a b s t r a c t Many animal phyla contain clades in which most or all species are venom-injecting predators. An example, in the arthropods, is the class Chilopoda, containing the approximately 3500 species of centipedes. Very little ecological or behavioural work yielding quantitative data has been conducted on centipede predation. Here, we describe a study of this kind. Our experiments employed one centipede species – a large tropical one, Scolopendra subspinipes mutilans – and two species of prey – a cricket, Gryllus assimilis, and a locust, Schistocerca gregaria. We conducted two experiments. The first was aimed at investigating the extent to which the centipedes attacked prey in particular tagmata as opposed to at random over the whole body surface. The results showed that the centipedes were highly selective, preferring to attack the head or thorax rather than the abdomen; indeed, they often reoriented the prey in order to achieve this. A possible explanation of this behaviour is to maximize the speed with which the neurotoxins in the venom reach either the brain or the thoracic ganglia that control limb movement. The second experiment was aimed at investigating the effect of venom-extraction on the attack rate, and specifically at testing if the magnitude of any such effect differed between the two types of prey, which differ considerably in size. The results showed a major effect of venom extraction in relation to both types of prey, but with the time taken to return to a ‘normal’ attack rate being longer in the case of the larger prey-type, namely the locust. We discuss these results in relation to the ‘venom optimization hypothesis’ and, more generally, to the principle of minimizing the production/use of venom, which is an energetically expensive resource. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The class Chilopoda (Arthropoda: Myriapoda) contains approximately 3500 described species of centipedes distributed among five living orders (Edgecombe and Giribet, 2007). Centipedes are predators characterised by a dorsoventrally flattened body bearing 15 or more pairs of legs, one pair per trunk segment. The head is attached to the trunk, which is not divided into thorax and abdomen. The first post-cephalic segment bears a pair of strong and sharp claws, named forcipules, each containing a venom system. Proximally, the venom-producing gland is encapsulated deep into the forcipule. Distally, the forcipule finishes into a sharp apical claw bearing a subterminal opening (the meatus) through which the venom is released (Fig. 1) (Dugon and Arthur, 2012). With the exception of a few species (e.g.: the scolopendromorph Paracryptops spinosus with unusually short forcipules – see Jangi and Dass, 1978), centipedes are thought to feed predominantly on live animals (Voigtländer, 2011). Centipedes seize their prey with the anterior legs and usually envenomate them with a stabbing ⇑ Corresponding author. E-mail address: [email protected] (M.M. Dugon). 0022-1910/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jinsphys.2012.03.014
motion of the forcipules. Plant material is also sometimes ingested but this intake is thought to be negligible (Lewis, 1981). Centipedes are considered opportunistic predators, catching any prey that can be overpowered. Large centipedes of the order Scolopendromorpha, the ‘giants’ of the centipede world, reaching up to 30 cm long, do sometimes attack large arthropods (scorpions; Iorio, 2004), amphibians (toads: Carpenter and Gillingham, 1984), reptiles (small snakes: Okeden, 1903 reported by Lewis, 1981), and also small mammals (mice: Voigtländer, 2011; bats: Molinari et al., 2005). Little is known about the techniques used by centipedes to capture prey items. Scolopendromorphs display a variety of hunting methods, from wandering on the leaf-litter, flicking the anterior half of the body on both sides in order to make contact with the prey and outrun it, to actively catching prey in mid-flight, sometimes hanging upside down from the ceiling of a cave (Molinari et al., 2005). The scolopendrid Scolopendra viridis has been reported to consistently use the anterior walking legs to hold the prey while venom is delivered with the forcipules (Elzinga, 1994). Several species have been observed seizing prey with the terminal pair of legs, and then flicking the anterior half of the body to inject the prey with venom.
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
875
Fig. 1. Left: Scolopendra sp. from Hubei, China, preening the antennae using the 2nd pair of maxillae. The forcipules are clearly visible as a pair of sharp claws curving anteromedially under the head capsule. Right: Scolopendra subspinipes mutilans attacking a cricket. The forcipules, having stabbed the prey, are holding it firmly in place while the venom takes effect.
Venom is a key element in the predatory behaviour of centipedes. Whether venom is solely used for immobilising prey by inducing paralysis remains unclear. It has been proposed that venom also contains digestive enzymes used to soften-up the flesh of the prey which is subsequently sucked-up (Undheim and King, 2011). The exoskeleton of the prey is usually not consumed. So far, venomic analyses have been performed only on large scolopendromorphs known to have a medically significant bite. Noda et al. (2001) isolated the benzoic alkaloid Scolopendrine as well as serotonic and muscarinic components, the latter two being powerful pain triggers and toxic catalysers. The neurotoxicity of Scolopendra sp. venom has been demonstrated by Gutierrez et al. (2003). Rates et al. (2007) and Malta et al. (2008) provided the most recent comparative study on centipede venom, based on South-American scolopendromorphs. The authors showed the myotoxic and weak haemolytic properties of the venoms. Their findings corroborate the observations of Bücherl (1946, 1971) on small mammals injected with diluted venom. Arthropods injected with venom are disabled within seconds (Voigtländer, 2011), a result expected in the presence of neurotoxins. The presence of myotoxic and haemolytic effects may be congruent with the hypothesis of external digestion. Undheim and King (2011) recently reviewed the literature pertaining to the composition and toxicity of centipede venoms. The behavioural processes involved in prey capture and venom injection in centipedes are virtually unknown, even in the largest species of scolopendromorphs. Physiological experiments show that stimulation of the forcipular sensilla may trigger a mechanical feeding response (Jangi and Dass, 1977). However, nothing is known about the role of venom availability in the general feeding response of centipedes. Here we investigate two aspects of predatory behaviour in the Chinese red-headed centipede Scolopendra subspinipes mutilans (Scolopendromorpha: Scolopendridae). The first experiment consisted of recording the behaviour of prey capture and evenomation. A second experiment was designed to investigate the relationship between predatory behaviour, venom availability and prey-type. 2. Materials and methods Two distinct experiments were performed: 1. A prey orientation experiment, in which 50 centipedes were presented each in turn with a prey item (the field cricket Gryllus assimilis). The reaction of each centipede was recorded and later analysed in order to assess if S. s. mutilans reorientate the prey after the initial seizure to inject venom in a ‘favoured’ part of the prey.
2. An experiment investigating prey attack in relation to venom availability. 150 S. s. mutilans were used in this experiment, in which the centipedes were offered one or another of two types of prey items ( G. assimilis or the migratory desert locust Schistocerca gregaria) at different times after venom extraction. 3. General set-up Prior to experiments 1 and 2, the centipedes were starved for 2 weeks. As centipedes are thought to be mainly nocturnal predators, each set of treatments (for both experiments 1 and 2) started at about 10 pm. The experimental ‘arena’ consisted of a translucent acrylic container (200 mm 140 mm 115 mm). No substrate was used and the container was briefly cleaned before introducing a new specimen into the arena. No artificial light was directed at the arena, and the observations were performed using dimmed, ceiling-fixed, fluorescent tubes. The experiments were performed using one centipede and one prey item at a time. Each centipede was introduced into the arena and was allowed to settle for 2 min. A prey item was then introduced into the enclosure containing the centipede. If the prey and the centipede were immobile, the prey was induced, with a gentle brush stroke, to walk towards the centipede. The behaviour of the centipede was recorded for 3 min. At the end of the 3 min, the centipede was removed from the experimental arena and returned to its normal container. 4. Prey items The field crickets and the desert locusts were purchased online from Livefood, UK. Large specimens of both prey species were selected for the experiments. From the base of the antennae to the end of the abdomen, the selected G. assimilis measured a minimum of 18 mm and a maximum of 22 mm (average: 20 mm). This resulted in a weight range between 400 and 500 mg (average: 438 mg). The selected specimens of S. gregaria measured at least 40 mm (average: 48 mm) and weighed between 1300 and 2300 mg (average: 1561 mg). Both prey items belong to the order Orthoptera and are considered to be similarly vulnerable to the venom of Scolopendra species (Iorio, 2004). However, as S. gregaria is much larger and heavier than G. assimilis, it was assumed that more venom would be necessary to tackle a S. gregaria. 4.1. Experiment 1: prey orientation experiment The goal of this experiment was to determine if centipedes manipulate their prey (in this case always a G. assimilis) in order to orientate them in a particular fashion prior to venom delivery. Consistency in the orientation of the prey and consistency in the
876
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
tagma into which venom is delivered may indicate that centipedes inject venom where it is most effective. 50 specimens of S. s. mutilans were tested in this experiment. Up to four behavioural events were recorded: Prey attacked/not attacked: A successful attack consisted of the prey being seized with the forcipules and not released during the 3 min of observation, regardless of subsequent manipulations. Prey seized and then released: Unsuccessful attacks occurred when the prey was seized and subsequently released within the 3 min of observation. Location of the initial bite: This consists of recording the part of the body of the prey into which the forcipules were first inserted during the seizure. Two body parts are distinguished on the prey: (1) the head + thorax; (2) the abdomen. Location of the final bite: This is the location of the forcipules at the end of the 3 min of observation. If manipulations occurred, the location of the final bite may be different from the initial bite. Only movements involving a change of bite location from head + thorax to abdomen or from abdomen to head + thorax were recorded as a changed location. If centipedes insert their forcipules anywhere indiscriminately on the prey, the expected frequencies for bite location (head + thorax or abdomen) should not be significantly different from the proportions these parts occupy on the prey’s body. These proportions were assessed by measuring the length of the head + thorax and the length of the abdomen in ten specimens, both on their dorsal and ventral sides. The values were then averaged. In the case of G. assimilis, the head + thorax represented about 37% of the whole body length, the abdomen representing the remaining 63%. Therefore, in the case of spatially random envenomation, 37% of bites were expected to be delivered on the anterior tagmata and 63% were expected to occur on the abdomen. The significance of the results was tested using either a chisquare goodness-of-fit test (Table 3) or a chi-square test for independence (Table 4). The analyses were performed using Graphpad (http://www.graphpad.com).
4.2. Experiment 2: venom availability and prey attack rate experiment This experiment was designed to test if the attack rate on two types of prey items differs significantly depending on the amount of venom available in the venom glands of the centipede. The experiment consisted of extracting venom and then giving either a small prey (the cricket) or a larger prey (the locust) to different centipedes at different times after venom extraction. If centipedes have the ability to ‘decide’ whether to attack prey depending on the venom available in the venom gland, crickets should be attacked sooner after venom extraction than locusts. In this experiment, 150 centipedes were divided into six groups of 25 specimens each, as listed below. Four groups of centipedes had their venom extracted, and two groups were used as negative control groups, one for each type of prey. Prior to venom extraction, the centipedes were anaesthetised by CO2 exposure. Then, each centipede was attached, ventral side facing up, to an acrylic sheet using two pieces of adhesive tape. Venom was extracted by electro-stimulation using a double pulse stimulator (Cat. 6032, Scientific and Research Instruments Ltd.). Each electrical impulse lasted for 5 ms at 22 V and recurred every 110 ms at the same voltage. The tips of the electrical probes were equipped with wet sponges regularly immersed in a saline solution. The sponges were applied to the sides of the forcipules for approximately 2 s. This operation was repeated 3–4 times in order to extract as much venom as possible. The venom was collected using a pipette between each repeat in order to clean the area and to confirm good extraction of the venom at each application
of the probes. The specimen was then removed from the sheet and replaced into its enclosure. Control groups assured that the feeding response was not affected by the anaesthesia and electric shocks. The specimens were treated as for a venom extraction but electric impulses were never applied on the forcipules. Instead, one electrical probe was applied on the third leg-bearing segment and the other on the twentieth segment so that although the current ran through most of the body, it did not pass through the anteriormost segments or the head capsule. Any application of the probes on the first or second leg-bearing segment would result in venom being discharged. The six groups of 25 centipedes each were treated as follows (see also Table 1): Group 1 (control): G. assimilis were given 6 h after anaesthesia and electrocution. Group 2: G. assimilis were given 6 h after anaesthesia and venom extraction. Group 3: G. assimilis were given 24 h after anaesthesia and venom extraction. Group 4 (control): S. gregaria were given 6 h after anaesthesia and electrocution. Group 5: S. gregaria were given 24 h after anaesthesia and venom extraction. Group 6: S. gregaria were given 48 h after anaesthesia and venom extraction. The prey-seizure frequencies were recorded for each group. The following between-group prey-seizing frequencies were compared in 2 2 contingency tables using a chi-square test with one degree of freedom: Group 1 (neg. ctrl/6 h/G. assimilis) versus group 2 (ven. ext./6 h/ G. assimilis). Group 2 (ven. ext./6 h/G. assimilis) versus group 3 (ven. ext./ 24 h/G. assimilis). Group 1 (neg. ctrl/6 h/G. assimilis) versus group 4 (neg. ctrl/6 h/ S. gregaria). Group 3 (ven. ext./24 h/G. assimilis) versus group 5 (ven. ext./ 24 h/S. gregaria). Group 4 (neg. ctrl/6 h/S. gregaria) versus group 5 (ven. ext./24 h/ S. gregaria). Group 5 (ven. ext./24 h/S. gregaria) versus group 6 (ven. ext./ 48 h/S. gregaria).
The results of experiment 2 were analysed using both chisquared and Fisher’s exact tests.
5. Results 5.1. Experiment 1: prey orientation prior to envenomation Of a total of 50 S. s. mutilans that were offered an adult G. assimilis, 49 (98%) seized the prey within the 3 min observation timeframe. A single centipede (2%) showed no interest in the cricket. None of the centipedes subsequently released the prey within the 3 min. Once seized, the prey was reoriented in 14 cases (29%) (Table 2). 28 (57%) of the 49 seizure events occurred on the head + thorax. The abdomen was seized in 21 cases (43%). As can be seen from table 3, these observed frequencies are significantly different from the expected ones, indicating that centipedes preferentially attack the head and thorax rather than the abdomen during the initial seizure.
877
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
Table 1 Treatment details for 6 groups of adult Scolopendra subspinipes mutilans in the venom availability experiment. Each group is composed of 25 specimens starved for 2 weeks prior to the start of the experiment. Group 1 Prey item Treatment Resting period before feeding
Group 2
Gryllus assimilis (cricket) Control Venom extraction 6h 6h
Group 3
Group 4
Venom extraction 24 h
Schistocerca gregaria (locust) Control Venom extraction 6h 24 h
Prey not attacked 1 Prey not released 49 Prey not reoriented 35
Group 6 Venom extraction 48 h
reorientate their prey in order to inject venom into the head + thorax rather than into the abdomen.
Table 2 Attack frequency of Scolopendra subspinipes mutilans on Gryllus assimilis. Prey attacked 49 Prey released 0 Prey reoriented 14
Group 5
Total 50 Total 49 Total 49
5.2. Experiment 2: prey attack rate and venom availability
The prey was not reoriented in 71% of successful attacks. Among these, 77% of bites occurred on the head + thorax of G. assimilis; the abdomen was seized in the remaining 23% of cases (Table 3). This demonstrates the preference of S. s. mutilans for biting the anterior tagmata of the prey. Reorientation occurred from head + thorax to the abdomen in 7% of all reorientations, and from the abdomen to the head + thorax in the remaining 93%. The expected frequencies of bite location (on the basis of random manipulation) were determined by the relative sizes of the head + thorax and abdomen in G. assimilis (Table 3). These expected frequencies were not confirmed, thus inferring that reorientation of the prey is significantly biased towards a secondary bite located on the head + thorax. At the end of the 3 min of observation, and regardless of reorientation, 82% of the final bites were located on the head + thorax and 18% were located on the abdomen. The high significance of the departure from the expected bite location frequencies (Table 3) demonstrates that centipedes do not bite the prey randomly but definitely favour venom injection in the head or thorax. The difference between the observed frequencies of initial and final bite locations was highly significant (Table 4). This is consistent with previous results and with the hypothesis that centipedes
5.2.1. G. assimilis as a prey item When venom was not extracted, 96% the centipedes attacked the prey. A single specimen (4%) disregarded the cricket. None of the centipedes that attacked subsequently released the prey. When the observed attack frequencies of the G. assimilis control group are compared to the observed attack frequencies in the prey orientation experiment, it can be seen that the anaesthesia and the electric impulses do not significantly change the predatory behaviour of the centipedes (Table 5A). Six hours after venom extraction, only 16% of the centipedes bit the prey-cricket and held onto it. Twenty-four percent of centipedes seized and then released the prey. The prey was totally ignored in the remaining 40% of cases. The ‘prey-seized-andreleased’ cases are treated as ‘prey not attacked’. When comparing group 1 (G. assimilis – Control) versus group 2 (G. assimilis – 6 h extraction), it appears that the frequency of successful attacks 6 h after venom extraction is significantly lower than the frequency of successful attacks when venom is not extracted (Table 5B). In the event of a ‘prey-seized-and-released’, most released prey crickets are capable of fleeing despite the stabbing wounds caused by the sharp apical claws of the centipede legs involved in seizing. This observation is consistent with the expected non-effectiveness of the venom apparatus 6 h after venom extraction. 24 h after venom extraction, 92% of centipedes belonging to the group 3 (G. assimilis – 24 h extraction) seized and held onto the prey. One specimen performed a ‘seize-and-release’ and another one ignored the prey completely. These results are significantly
Table 3 Expected and observed frequencies of initial seizure and envenomination of a prey cricket Gryllus assimilis by Scolopendra subspinipes mutilans. Out of 50 trials, initial seizure of the prey by the centipede occurred 49 times. In 35 cases, the prey was subsequently envenomated without any further manipulation. Reorientation of the prey occurred in 14 cases. The final envenomation indicates the location of the definitive bite at the end of the 3 min of observation. Expected frequencies were calculated on the basis of random seizure or manipulation, proportionally to the relative size of the tagmata in relation to whole body length.
Initial seizure (N = 49) Envenomation location, no reorientation (N = 35) Reorientation + envenomation (N = 14) Final envenomation location (N = 49)
Location
Expected frequencies
Observed frequencies
X2 Goodness-of-fit
Head + thorax Abdomen Head + thorax Abdomen From head + thorax to abdomen From abdomen to head + thorax Head + thorax Abdomen
18 31 13 22 9 5 18 31
28 21 27 8 1 13 40 9
X2 = 8.781 p = 0.003 X2 = 23.986 p < 0.0001 X2 = 19.119 p < 0.0001 X2 = 42.502 p < 0.0001
Table 4 Observed frequencies of the initial and final bite locations following successful attacks on Gryllus assimilis by Scolopendra subspinipes mutilans.
Initial bites Final bites
Bites located on the head + thorax
Bites located on the abdomen
Total
Chi square for independence test
28 40
21 9
49 49
X2 = 6.918 p < 0.01
878
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
Table 5 Observed attack frequencies of Scolopendra subspinipes mutilans on two types of prey items at different times after venom extraction. Group 1: control. Gryllus assimilis given 6 h after anaesthesia and electric shocks. Group 2: Gryllus assimilis given 6 h after venom extraction. Group 3: Gryllus assimilis given 24 h after venom extraction. Group 4: control. Schistocerca gregaria given 6 h after anaesthesia and electric shocks. Group 5: Schistocerca gregaria given 24 h after venom extraction. Group 6: Schistocerca given 48 h after venom extraction. Prey attacked
Prey not attacked
Total
X2
P
1
50
0.26
ns
32.47
p < 0.0001
29.07
p < 0.0001
35.28
p < 0.0001
13.23
p < 0.001
0.36
ns
35.28
p < 0.0001
A
Experiment 1 Vs. Group 1 (G.assimilis – Control)
49 24
1
25
B
Group 1 (G. assimilis – Control) Vs. Group 2 (G. assimilis – 6 h extraction)
24
1
25
4
21
25
C
Group 2 (G. assimilis – 6 h extraction) Vs. Group 3(G. assimilis – 24 h extraction)
4
21
25
23
2
25
D
Group 4 (S. gregaria – Control) Vs. Group 5 (S. gregaria – 6 h extraction)
23
2
25
2
23
25
E
Group 5 (S. gregaria – 24 h extraction) Vs. Group 6(S. gregaria – 48 h extraction)
2
23
25
14
11
25
F
Group 4 (S. gregaria – Control) Vs. Group 1 (G.assimilis – Control)
23
2
25
24
1
25
G
Group 3 (G. assimilis – 24 h extraction) Vs. Group 5 (S. gregaria – 24 h extraction)
23
2
25
2
23
25
different from the results obtained 6 h after venom extraction (group 2), as shown in Table 5C. The results are very similar to those obtained with the control group, when venom was not extracted.
6. Discussion
5.2.2. Schistocerca gregaria as a prey item When no venom was extracted (group 4), 92% of centipedes attacked the locust 6 h after electrocution and 8% showed no interest in it. None of the centipedes that made an attack subsequently released the prey. Twenty-four hours after venom extraction, only 8% of centipedes seized and held their prey. In 36% of cases, the prey was seized and then released. Interestingly, the ‘seized-and-released’ locusts appeared (at least superficially) unharmed, and were capable of retaliation by ‘kicking’ the centipedes with the posterior legs. Fifty-six percent of centipedes ignored the prey. If the ‘prey-seizedand-released’ cases are treated as ‘prey not attacked’ and the results of this (group 5) are tested against the results of the control (group 4), it can be seen that the frequency of successful attacks 24 h after venom extraction is significantly lower than when venom is not extracted (Table 5D). Forty-eight hours after venom extraction (group 6), 56% of centipedes seized and held onto their prey; 24% performed a ‘seizeand-release’. Twenty percent of centipedes did not attack the prey at all. The frequency of successful attacks significantly increased at 48 h compared to the frequency at 24 h after venom extraction (Table 5E). However, the number of successful attacks was not at the level seen in the control group in which venom was not extracted.
When seizing a prey cricket, a centipede usually bites its head and thorax rather than its abdomen. A bite located on the abdomen is usually followed by a reorientation of the prey and by a secondary bite on the head or thorax. The number of final bites on the anterior tagmata is significantly higher than the expected number. This observation leads to two questions. First, why would centipedes show a preference for injecting venom in the anterior end of the prey rather than the posterior? And second, how do centipedes spatially comprehend both their own position and the position of the prey? The main role of venom is to disable prey in an effective and prompt manner (the second role being the lysis of internal tissues). The venoms of Scolopendridae are known to contain, among many other elements, neurotoxins disrupting synaptic signals (Stankiewicz et al., 1999; Gutierrez et al., 2003). We can suspect that these neurotoxins are the major fast-acting disabling elements of many centipedes’ venoms. By changing the properties of the synaptic channels, neurotoxins may act as strong paralytics when reaching the central nervous system. Therefore, the venom is most efficient when injected either into the head capsule or close to the thoracic ganglia controlling locomotion. In the case of insects these ganglia are present on the ventral side of the thorax. When injected into the soft tissues of the abdomen, it is likely that the venom would be diluted in the circulatory system and the gut of the prey, and would have to travel further in order to disable locomotion and prevent the escape of or retaliation by the prey. However, considering the amount and the potency of the venom delivered by large scolopendromorphs (see Bücherl, 1946, 1971), it is very likely that the location of venom injection does not matter much in order to disable a small or average-sized prey item. Prey orientation may be based on two complementary principles. The first is a principle of economy: venom is physiologically expensive to produce and the replenishment of the venom glands requires time. The second principle is based on limiting prey retaliation: seizing the prey with the walking legs in the alignment of
5.2.3. Between prey-type analysis When the results of the control group with G. assimilis preytype (group 1) are tested against the control group with S. gregaria prey-type (group 4), they are not significantly different (Table 5F). The anaesthesia and the electric impulses do not significantly change the predatory behaviour of the centipedes, and S. s. mutilans attacks both prey types at the same rate. A comparison of the attack frequencies between centipede groups that were given G. assimilis (group 3) or S. gregaria (group 5) 24 h after venom extraction showed that centipedes attack the crickets more than the locusts (Table 5G).
6.1. Prey orientation experiment
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
the anterior-posterior body axis while stabbing its thorax or head with the forcipules restrains the prey sufficiently to avoid retaliation until the venom takes effect. The results of the prey orientation experiment also provide some information on the capacity of centipedes to perceive the three-dimensionality of their surroundings. Because the primary attack occurs within a few tenths of a second, it is unlikely that the centipede always manages to ‘understand’ the spatial orientation of the prey through physical contact with the antennae or another body part prior to the attack. Spatial orientation must also occur through other means. First, it is probable that the perception of the prey’s orientation takes place once the centipede has pounced on the prey and ‘felt’ the prey with its legs and immobilised it, prior to any forcipular movement. This would explain the occurrence of pre-bite manipulations and the systematic use of the walking legs during the grasping motion. Second, we could consider that, to a limited extent, S. s. mutilans uses its eyesight at very short distances to perceive a prey’s orientation before the attack. However, a specific role of eyesight in prey capture by scolopendromorphs is unlikely in view of previous reports on the subject, which all suggest contact, mechanoreception or chemoreception as the main tools for prey recognition and localisation (Cloudsley-Thompson, 1955 mentioned by Lewis, 1981; Manton, 1965). 6.2. Venom availability and prey attack rate experiment Overall, the results observed in experiment 2 demonstrate two distinct patterns: First, it appears that the hunting behaviour of centipedes is linked to venom available in the glands. The behavioural pattern is altered in two ways by reduced venom: (1) a large proportion of centipedes refuse to attack the prey; and (2) some centipedes release their prey after seizing it, a phenomenon not observed when there was no venom extraction. The physiological mechanisms allowing centipedes to perceive the amount of venom available and subsequently to change their behaviour is still unknown. However, the occurrence of failed attacks suggests also that centipedes may sometimes ‘override’ the physiological feedback and engage themselves in an attempt to attack a prey. Because the centipedes used in these experiments were starved for 2 weeks, it is possible that hunger prevailed over lack of venom availability and the risks linked to retaliation by the prey. Also, it should be noted that centipedes involved in failed attacks released their prey after a few seconds. It is possible that the mechanoreceptors and chemoreceptors located on the tarsungulum (distal end) of the forcipules provide feedback on the amount of venom discharged and the condition of the prey in the seconds following the bite. In Scolopendra morsitans, the application of a glucose solution on the sensilla of the tarsungulum triggers a response of the feeding apparatus, presumably acting via chemoreceptors (Jangi and Dass, 1977). The second distinct pattern is the time lapse between venom extraction and the return towards a ‘normal’ hunting response. When the prey is small (G. assimilis, with an average weight of 438 mg), it seems that centipedes have produced enough venom to tackle the prey normally 24 h after venom extraction. When the prey is much larger (S. gregaria, with an average weight of 1561 mg), most centipedes refuse to attack the prey 24 h after venom extraction. It is only 48 h after venom extraction that the frequency of successful attacks increases significantly and even then it has not returned to its ‘normal’ level. These observations demonstrate that centipedes do perceive the difference between small and large prey, and alter their behaviour accordingly. The amount of venom available, the size of the prey and possibly the resistance of the prey to the venom may all be taken into account before any ‘decision’ is made about attacking. But once again the number of
879
failed attacks tempers this idea, as wrong ‘decisions’ lead sometimes to unsuccessful attempts, a phenomenon not observed when the venom gland is full. Venom-economy strategies by means of selective venom injection and prey orientation have been previously observed and investigated only in very few species of terrestrial arthropods, such as the wandering spider Cupiennius salei (Wullschleger and Nentwig, 2002; Wigger et al., 2002; Kuhn-Nentwig et al., 2004; Hostettler and Nentwig, 2006; Boevé, 1994), the New-Zealand crab spider Diaea sp. (Pollard, 1990), the scorpion Parabuthus transvaalicus (Nisani and Hayes, 2011), the arboreal African ponerine ant Platythyrea conradti (Dejean, 2011) and a predatory wasp (Haspel et al., 2003). Studies have shown that P. conradti turns its prey over and stings them almost exclusively on the ventral part of the thorax, where the neural cord is closest to the exoskeleton. Also, P. conradti injects paralytic venom only into large prey presenting a retaliation potential (Dejean, 2011). Some predatory wasps inject venom specifically into the brain of their prey (Haspel et al., 2003). This principle of venom economy has been named the venom optimization hypothesis (Wigger et al., 2002). In the case of S. s. mutilans (and P. conradti for that matter), the venom optimization occurs in two ways: (1) orientate the prey in order to deliver venom in the most sensitive area; and (2) attack a prey matching the amount of venom available in the gland. Interestingly, the few species in which venom optimisation has been observed belong to three different subphyla of Arthropoda (Chelicerata, Hexapoda and now Myriapoda). Given the diverse scattering of major venomous groups (like centipedes) across the arthropod evolutionary tree, it seems likely that venom optimization has evolved convergently several times. In each case, the basis of the evolutionary trend may be the principle of economising venom, an expensive resource. Acknowledgements We thank Eoin MacLoughlin and ‘Uncle’ Daniel Leong for logistic support. This work was funded by a Professor Brendan F. Keegan memorial Ph.D. studentship (held by M.D.) from Eli Lilly, to whom we are grateful for their financial support. References Boevé, J.L., 1994. Injection of venom into an insect prey by the free hunting spider Cupiennius salei (Araneae, Ctenidae). Journal of Zoology 234, 165–175. Bücherl, W., 1946. Açâo do veneno dos escolopendromorfos do Brasil sobre alguns animais de laboratorio. Memórias do Instituto Butantan 19, 181–197. Bücherl, W., 1971. Venomous chilopods or centipedes. In: Bücherl, W. and Buckley, E.E. (Eds.), Venomous Animals and Their Venoms. Academic Press, New-York, pp. 169–196. Carpenter, C.C., Gillingham, J.C., 1984. Giant centipede (Scolopendra alternans) attacks marine toad (Bufo marinus). Caribbean Journal of Sciences 20, 71–72. Cloudsley-Thompson, J.L., 1955. Some aspects of the biology of centipedes and scorpions. The Naturalist 6, 147–153. Dejean, A., 2011. Prey capture behavior in an arboreal African ponerine ant. Public Library of Science One 6, e19837. Dugon, M.M., Arthur, A., 2012. Comparative studies on the structure and development of the venom-delivery system of centipedes; and a hypothesis on the origin of this evolutionary novelty. Evolution & Development 14, 1– 10. Edgecombe, G.D., Giribet, G., 2007. Evolutionary biology of centipedes (Myriapoda: Chilopoda). Annual Review of Entomology 52, 151–170. Elzinga, R.J., 1994. The use of legs as grasping structures during prey capture and feeding by the centipede Scolopendra viridis Say (Chilopoda: Scolopendridae). Journal of the Kansas Entomological Society 67, 369–372. Gutierrez, M.C., Abarca, C., Possani, L.D., 2003. A toxic fraction from Scolopendra venom increases the basal release of neurotransmitters in the ventral ganglia of crustaceans. Comparative Biochemistry and Physiology 135, 205–214. Haspel, G., Rosenberg, L.A., Libersat, F., 2003. Direct injection of venom by a predatory wasp into cockroach brain. Journal of Neurobiology 56, 287–292. Hostettler, S., Nentwig, W., 2006. Olfactory information saves venom during preycapture of the hunting spider Cupiennius salei (Araneae: Ctenidae). Functional Ecology 20, 369–375.
880
M.M. Dugon, W. Arthur / Journal of Insect Physiology 58 (2012) 874–880
Iorio, E., 2004. Les appareils venimeux des Chilopodes: mécanismes et pathologies. Bulletin de Phyllie 20, 23–33. Jangi, B.S., Dass, C.M.S., 1977. Chemoreceptive function of the poison fang in the centipede Scolopendra morsitans Linn. Indian Journal of Experimental Biology 15, 803–804. Jangi, B.S., Dass, C.M.S., 1978. A new species of the genus Paracryptops from India, with remarks on the generic and specific characters (Chilopoda: Scolopendromorpha: Cryptopidae). Zoological Journal of the Linnean Society 64, 327–330. Kuhn-Nentwig, L., Schaller, J., Nentwig, W., 2004. Biochemistry, toxicology and ecology of the venom of the spider Cupiennius salei (Ctenidae). Toxicon 43, 543– 553. Lewis, J.G.E., 1981. The Biology of Centipedes. Cambridge University Press, UK. Malta, M.B., Lira, M.S., Soares, S.L., Rocha, G.C., Knysak, I., Martins, R., Guizze, S.P.G., Santro, M.L., Barbaro, K.C., 2008. Toxic activities of Brazilian centipede venoms. Toxicon 52, 255–263. Manton, S.M., 1965. The evolution of arthropodan locomotory mechanisms. Part 8. Functional requirements and body design in Chilopoda. Journal of the Linnean Society of London 46, 251–484. Molinari, J., Gutierrez, E.E., De Ascencao, J.M.N., Arends, A., Marquez, R.J., 2005. Predation by giant centipedes, Scolopendra gigantea, on three species of bats in a Venezuelan cave. Caribbean Journal of Sciences 41, 340–346. Nisani, Z., Hayes, W.K., 2011. Defensive stinging by Parabuthus transvaalicus scorpions: risk assessment and venom metering. Animal Behaviour 81, 627– 633.
Noda, N., Yashiki, Y., Nakatani, T., Miyahara, K., Du, X.-M., 2001. A novel quinoline alkaloid possessing a 7-benzyl group from the centipede, Scolopendra subspinipes. Chemical & Pharmaceutical Bulletin 49, 930–931. Okeden, W.P., 1903. A centipede eating a snake. Journal of Bombay Natural History Society 15, 1. Pollard, S.D., 1990. The feeding strategy of a crab spider, Diaea sp. indet. (Araneae: Thoisidae): post-capture decision rules. Journal of Zoology 222, 601–615. Rates, B., Bemquerer, M.P., Richardson, M., Borges, M.H., Morales, R.A.V., De Lima, M.E., Pimenta, A.M.C., 2007. Venomic analyses of Scolopendra viridicornis nigra and Scolopendra angulata (Centipede, Scolopendromorpha): shedding light on venoms from a neglected group. Toxicon 49, 810–826. Stankiewicz, M., Hamon, A., Benkhalifa, R., Kadziela, W., Hue, B., Lucas, S., Mebs, D., Pelhate, M., 1999. Effects of a centipede venom fraction on insect nervous system, a native Xenopus oocyte receptor and on an expressed Drosophila muscarinic receptor. Toxicon 37, 1431–1445. Undheim, E.A.B., King, G.F., 2011. On the venom system of centipedes (Chilopoda), a neglected group of venomous animals. Toxicon 57, 512–524. Voigtländer, K., 2011. Chilopoda – Ecology. In: Minelli, A. (Ed.), The Myriapoda, vol. 1. Brill, Leiden, pp. 309–325. Wigger, E., Kuhn-Nentwig, L., Nentwig, W., 2002. The venom optimisation hypothesis: a spider injects large venom quantities only into difficult prey types. Toxicon 40, 749–752. Wullschleger, B., Nentwig, W., 2002. Influence of venom availability on a spider’s prey-choice behaviour. Functional Ecology 16, 802–807.