Remote monitoring of web vibration used by orb weaver spiders that run a signal thread from the web hub to a silken retreat is an alternative strategy to on-web monitoring (Eberhard 1990; Gregoric et al. 2015). Here, we presented data to measure the principal costs and benefits of using a signal thread for information-acquisition via web vibration. We found that using a signal thread did not increase the propagation time through the web, but it decreased the amplitude of propagating vibrations, and that mass present at the hub affected the timing and amplitude of vibrations. These findings have implications for the biological information that can be gained through each information-acquisition strategy, particularly the vibration source location.
Cost of a signal thread?
The signal thread off-hub strategy was not accompanied by a time cost of transmission to the spider; the time taken for transverse vibrations to propagate to the spider location either at the hub or on a signal thread was within a similar time range (greatest time difference was 5 ms). The mechanism controlling this is a wave speed gradient within radial threads, which is explained by the presence of the capture spiral, which increases the tension on the radial threads as they diverge (Wirth and Barth 1992; Mortimer et al. 2016). The webs of the two species were similar as they have similar tension gradients, where the signal thread is tensioned by Zygiella (Wirth and Barth 1992; Mortimer et al. 2015). The presence of a mass created bigger time differences between the two webs: whereas arrival times were slower for Zygiella, as there was no extra tensioning on the Zygiella web, the peak times were faster, as there was less of an inertial effect on the Zygiella web. But differences between the webs remained under 5 ms, which is small compared to the time needed to behaviourally react to the information. The shortest, i.e. quickest, time that either Zygiella or Araneus respond to a web vibration has been measured as 100 ms (Klärner and Barth 1982).
Attenuation, or loss of energy, represented an obvious cost of employing a signal thread—vibrations became more damped as they travelled further through silk. This attenuation cost was not as large for transverse waves when mass was present on the webs. Vibrations coming from the capture area opposite the signal thread were attenuated more than vibrations on radial threads closer to the signal thread. This was due to loss of energy as transverse waves propagating through the hub, also evident for Araneus webs, where stiff cross threads were present (Zschokke and Vollrath 1995). For Zygiella, this suggests that certain radials coupled with the signal thread better, i.e. with less energy loss, than others. Adjustment of radial thread coupling to the signal thread could be one method that Zygiella can use to alter the ability of the web to transmit vibrational information.
The signal thread attenuated vibrations in a frequency range under 600 Hz, but it remains to be seen whether this influences the ability of Zygiella to discriminate between vibration sources. Frequency-dependent attenuation is likely to change non-linearly for larger-amplitude vibration sources, as damping due to air drag and internal silk damping will increase at higher amplitudes (Kolsky 1964; Sensenig et al. 2012).
Overall, the signal thread attenuates vibrations, particularly those under 600 Hz, but it does not appear to have a cost in terms of the propagation time of vibrations. A spider using a signal thread is therefore likely to require lower amplitude detection thresholds or prey-generated vibrations of higher magnitude to achieve similar prey capture success. This has implications for how the two spiders respond to vibrations, given that time, frequency and amplitude can be used as information to determine the location of the vibration source.
Locating the vibration source
The location of a vibration source is an important piece of information encoded within web vibration. In theory, time, amplitude or frequency differences between legs or different wave types can be used to determine the vibration source location (Mortimer 2017).
Starting with the timing component, the smallest difference between onset times of two vibrations that the spiders can sense with their slit sensilla has been recorded to be between 2 and 4 ms (Hergenröder and Barth 1983; Barth 1993). We found that the pre-stress gradients in the web resulted in propagation time differences that were so small between different input locations that the spider would not be able to detect any differences. Conversely, if there was no tension gradient along a radial (i.e. a constant low wave speed of 40 m s−1), the maximum possible time delay for these web geometries would be 4.4 and 7.2 ms to the hub (17.6 cm) and signal thread (28.8 cm) respectively. The capture spiral is therefore not only important for prey retention (Foelix 2010), but also has significant implications for vibration transmission as it affects pre-stress of the radials (Mortimer et al. 2016).
The fast speeds of both longitudinal and transverse waves would imply that it is highly unlikely that the arrival time can be used for determining the vibration source location, whether at the hub or on the signal thread, as propagation time differences between transverse and longitudinal waves are too short to be determined as separate arrival events (Hergenröder and Barth 1983; Barth 1993). Interestingly, placing a mass at the hub slowed down the arrival time of the peak amplitude of transverse waves sufficiently so that (i) the difference in peak time of longitudinal and transverse waves changed as a function of vibration input location and (ii) the difference in peak time between the locations was greater than the time detection threshold (2 ms). This observation requires further study, as transverse and longitudinal wave peak times may be useful information for prey localisation—when the spider is present at the hub.
Vibration amplitude is likely to be a richer source of information than vibration speed. Longitudinal waves are known to be directional cues (Masters and Markl 1981; Masters 1984a; Landolfa and Barth 1996), but it is unknown whether spiders can determine vibration source distance. For Zygiella, both transverse and longitudinal waves are likely to be important for triggering movement out of the retreat onto the hub, where sufficient amplitude is a possible trigger for a predatory response (Liesenfeld 1956; Klärner and Barth 1982; Masters 1984b). We note that Zygiella often orientates upon arrival at the hub (Klärner and Barth 1982). However, signal thread vibration may still encode locational information. This is a topic worth further study given the potential applications of vibration-sensing technologies (Tiwana et al. 2012; Fratzl and Barth 2009).
The observed patterns of vibration transmission highlight how their speed (time differentials) as well as amplitude and frequency could potentially be used by spiders to assess different sources of information in order to maximise prey capture success. The fast propagation speed of both wave types within the web would suggest that spiders are limited by the speed of their physiological processing of vibrational information rather than by the speed of physical propagation of wave information in the web (Klärner and Barth 1982). Longitudinal wave amplitude would provide information for orientation at the hub (Masters and Markl 1981; Masters 1984a), and differences in peak amplitude time and, more likely, amplitude between transverse and longitudinal waves could be used for determining distance to a vibration source. And as Masters (1984b) and Landolfa and Barth (1996) have postulated, spiders should be able to discriminate vibrational signals to inform decision making using both frequency and amplitude. The ability to accurately locate and discriminate vibration sources is likely to have fitness consequences for the spider, and future studies should compare Araneus and Zygiella in this respect to quantify the selection-relevant fitness costs for Zygiella when employing a signal thread.
As has become apparent, many factors determine the costs and benefits of employing the remote information-acquisition strategy of the off-hub spider such as Zygiella, where the vibrational information transfer is the key component. Our study highlights how wave propagation is affected by the delicate balance between physical limitations, e.g. attenuation, and factors under the spiders’ control, e.g. web geometry and radial thread tensioning. This shows how the spider’s web is a fascinating example of a biological structure engineered by evolution to enable the spider to balance the many benefits and costs of its extended phenotype.