Back to Insect page Previous page Next page
References *** SUMMARY of conclusions reached

Wilcox (2002) has uncovered evidence of sophisticated mimicry and deception among jumping spiders. I shall quote the abstract in full, as the evidence is so impressive:

Jumping spiders of the Genus Portia are aggressive mimics which prey primarily on other species of spiders, especially web-building spiders. The basic tactics Portia exhibits are invasion of a web and signaling deceitfully on it as if Portia were struggling insect prey, to attract the resident spider closer, wherepon Portia kills and eats the resident. In addition to such basic stalking motions on the web, Portia opportunistically uses background noise on the web to mask its stalking motions on the web; uses trial-and-error learning to choose an appropriate signal for a particular prey spider, and to determine a different ploy which may deceive the prey spider better; and makes detours which involve evaluation of a situation, planning ahead to execute a pathway which may take it initially away from and out of sight of the prey spider, and executing the detour, using a cognitive map, during a period of at least two hours. The flexible learning, planning ahead, and persistent maintenance of cognitive map abilities of Portia place it among the foremost of invertebrate groups in the cognitive arena.

The question of whether the spider is engaging in deception proper, with an intent to induce a false belief in another individual's mind, will be deferred until chapter 4. The important point is that the behaviour described here qualifies as mentalistic. If we examine the key features of operant conditioning, most or all seem to be present:

a drive (to prey on other spiders);

motor programs, an action selection mechanism and fine-tuning behaviour (the spider engages in trial-and-error learning or which allows it to fine-tune its body movements on the web, so that they match those of a struggling insect);

internal representations of its current motor output or efferent copy, and its afferent sensory inputs (not directly verified by Wilcox, but there is good reason to suppose this ability is present, given the fine-tuning behaviour described above);

associations between different motor commands and their consequences, which are stored in the animal's memory and updated when circumstances change (the spider "uses trial-and-error learning to choose an appropriate signal for a particular prey spider, and to determine a different ploy which may deceive the prey spider better");

a goal or end-state (getting close enough to the host spider to eat it);

a pathway for getting there (the detour, encoded in the spider's spatial memory);

sensory inputs (the spider can see and feel its way around the web);

a temporal correlation mechanism (without this, the spider would be unable to identify the right signal to attract the host spider); and

self-correction (the spider adjusts its motor behaviour to suit its host).

We can also speak of the spider as having beliefs - e.g. that its prey lies at the end of the path it is following - and desires (to eat other spiders).

In short, the behaviour described here is perhaps the clearest example of mentalistic behaviour examined so far in this thesis.