Field observations have shown that the ultimate legs of scolopendrids are used as hooks from which the animal can hang. Thus Remington (1950) reported that S. subspinipes Leach, 1815, fastened onto a tent near the ventilator hole with their ultimate legs and swung their bodies to one side or the other to seize insects that had alighted nearby. On 4.xi.95 Peter Daszak and Janet Cottingham found a female Scolopendra abnormis Lewis & Daszak, 1996, hanging by her ultimate legs in a hollowed out cavity beneath a slab of tuff on Serpent Island, Mauritius. The anterior part of the body was curled around 40 oval yellow eggs (Lewis unpublished data).

Molinari et al (2005) observed two specimens of Scolopendra gigantea L., 1758, feeding on bats whilst hanging from the roof of a Venezuelan cave to which they anchored using the last five or eight pairs of legs.

Kronmüller (2013, unpublished data) reported and photographed a Scolopendra dehaani that escaped the plastic container in the laboratory and was later found hanging on the camera tripod just using the ultimate legs (Figure 3A).

Warning posture

Lewis (unpublished data) noted in Nigeria that a small Scolopendra morsitans L., 1758, raised and splayed its ultimate legs when approached from behind by a large Mecistocephalus (Geophilomorpha). Neck (1985) described such threat display in Scolopendra viridis Gervais, 1847, the ultimate legs being spread apart and in Scolopendra heros Girard, 1853, where the legs were raised and waved back and forth, active defence involving biting or “holding” with the ultimate legs. A photo of an Ethmostigmus sp. sent to one of us (J G E L) by Chris Pennington from Krambach, New South Wales, Australia, showed the ultimate legs lifted and splayed revealing the spines on the median surface of the prefemora and on the coxopleural processes. A predator approaching from behind would come into contact with a battery of spines.

Autotomy and sound production

As noted above, in case of danger the members of the genus Alipes swing the ultimate legs from side to side and stridulate perhaps to distract or frighten potential predators.. They also use autotomy to detract potential predators like birds, lizards or mammals. In case of an attack, the centipede starts to stridulate which directs the attention of the predator to the ultimate legs which can be autotomised to make the escape easier. When detached with forceps the leaf-like ultimate legs of an Alipes grandidieri from Tanzania continued to stridulate for more than half a minute.

Cloudsley-Thompson’s (1961) observed that the long ultimate legs of the West African Rhysida nuda togoensis (now R. immarginata togoensis) slowly bend and straighten when detached and emit a faint creaking sound. This observations were repeated in Northern Nigeria in 1967 by Lewis (unpublished data): an ultimate leg of the same species detached with forceps produced very quiet squeaking for about 75 seconds, flexing at 2–3 second intervals and Kronmüller (2009, unpublished data) observed that the long slender ultimate legs of Rhysida immarginata immarginata (Porat, 1876) from the Philippines (Figure 1D) as well as Rhysida longipes (Newport, 1845) from various countries in Africa bend and straighten for a few minutes presumably to enable the centipede to flee while potential predators become distracted. In our experiments, when the ultimate legs were detached with forceps,they did not squeak as did those of Rhysida immarginata togoensis.

Video clips of Alipes sp. showing the typical defence behaviour including the sound of the stridulation as well as the bending movements of the detached ultimate legs of Rhysida sp. can be seen on the website http://www.scolopendromorpha.com.

Laboratory experiments were carried out to further investigate behaviour in prey capture as well as possible defence reactions in case of a predator attack. To simulate a the attack of mammal or bird predator, a piece of foam held with 25cm forceps was used to tap the centipedes from above.

The centipedes that were used for the experiments were either adult wild caught specimens (Scolopendra heros, Scolopendra subspinipes, Ethmostigmus trigonopodus, Scolopendra morsitans, Scolopendra hainanum, Cormocephalus aurantiipes) or, in one case, an adult captive bred specimen (Scolopendra galapagoensis).

The specimens were kept for at least 12 weeks before the experiments started in plastic containers of different sizes depending on the size of the specimen, with an air temperature of 25 degrees Celsius in day-time (around 10 hours illumination with a neon lamp) and 20–22 degrees Celsius at night-time. The average air humidity was around 70%. All plastic containers had a layer of humus and a piece of bark as a hiding-place. Once a week, the humus was sprayed with water to keep it moist and the specimen was fed one adult cricket (Gryllus assimilis).

For the experiments, the centipedes were moved to a 20 × 20 cm plastic container with a layer of about 2cm humus and left for at least one hour without disturbance before the experiments started. After each experiment, the centipedes were left for a further hour to recover before the next experiment.

The first experiment involved tapping the centipedes at different parts along the length of the body (anterior third, middle third and posterior third) to observe possible differences in defense reactions and to compare the responses of the investigated species.

In a second experiment a piece of bark was placed in the container as a refuge and the layer of humus was increased to 10cm. A centipede was placed in the container and as soon as the first third of the trunk was hidden either by the centipede crawling under the bark or by it digging into the humus, the tapping experiments described above were repeated.

Photos and short video clips are made with a Canon EOS 60D and a Canon EOS 6D, connected either to a Canon standard lens (18-55mm) or a Canon macro lens (100mm), three Canon Speedlight EX550/EX430 flashes (photos only), a tripod and a remote control (video clips).

Initial experiments were carried out on an adult Ethmostigmus trigonopodus (Leach, 1817) from Tanzania, length 13 cm. When tapped on the first third of the trunk (near the head), the centipede attacked the forceps with the forcipules. When tapped on the posterior third, or even at the ultimate legs, it adopted the warning position, raising the ultimate legs to display the ventral and medial prefemoral spines as well as the coxopleural processes which are also equipped with spines. In the most cases the centipede directed the warning position towards the forceps by turning the posterior segments. In some cases, depending on the strength and speed of the simulated attack, the centipede counter-attacked the forceps with the claws of the ultimate legs by chopping down on them after lifting them high into the warning position.

When the E. trigonopodus was tapped in the mid part of the trunk, the centipede curled sideways to reach the forceps with the forcipules and the ultimate legs simultaneously. These observations were repeated in experiments with Scolopendra heros var. arizonensis Kraepelin, 1903 (USA), Scolopendra galapagoensis (Peru), Scolopendra morsitans (Pakistan), Scolopendra cingulata (Spain), Scolopendra subspinipes (Indonesia), Scolopendra hainanum Kronmüller, 2012 (China), Scolopendra spinosissima Kraepelin, 1903 (Philippines), and Cormocephalus aurantiipes (Newport, 1844) (Australia). All showed the same warning posture raising the ultimate legs (Figure 2A).

In addition Scolopendra galapagoensis lifted not only the ultimate legs into the warning position but also the last 3-4 pairs of locomotory legs (Figure 2B). This may be to present the distodorsal prefemoral spines on these legs. This behaviour was not seen in S. heros var. arizonensis, the other New World species investigated which lacks these spines.

In the experiments with bark and a thicker layer of humus, no difference was observed in the behaviour depending on whether the specimen was tapped on the mid part of the trunk or on the last third. The result was the same when the centipede was fed (Schistocerca sp.) and the forcipules as well as the anterior pairs of locomotory legs were involved in feeding. In this situation the centipede could not use the forcipules to carry out a warning bite and the only defence behaviour was hitting with the ultimate legs and stabbing with the ultimate leg pretarsi (claws). The experiment was repeated with Ethmostigmus trigonopodus, Scolopendra galapagoensis, Cormocephalus aurantiipes and Scolopendra subspinipes, which all showed the same results.

Some scolopendrid centipedes, especially the giant Scolopendra sp. have striking colour patterns. In several species, the colouration of the last trunk segments and the ultimate legs is the same as the colouration of the head and antennae (Figure 3 B and C). Predators may not be able to differentiate between the head and the posterior part of the trunk.

Figure 3. 

A Scolopendra dehaani hanging from a camera tripod just using the ultimate legs B Scolopendra multidens and C Scolopendra heros showing possible auto mimicry as the ultimate legs and last segments of the trunk mimic the head and antennae.

The warning posture, especially that of Scolopendra galapagoensis, where the last three or four pairs of locomotory legs are raised in addition to the ultimate pair, resembles that of spiders such as Phoneutria sp. or tarantulas and may be a case of possible signal co-evolution.