Belt considers that the advantages gained by ant-mimicking Central American spiders lies entirely on the side of protection. In relation to this subject he says: “Ant-like spiders have been noticed throughout tropical America and also in Africa. The use that the deceptive resemblance is to them has been explained to be the facility it affords them for approaching ants on which they prey. I am convinced that this explanation is incorrect, so far as the Central American species are concerned. Ants, and especially the stinging species are, so far as my experience goes, not preyed upon by any other insects. No disguise need be adopted to approach them, as they are so bold that they are more likely to attack a spider than a spider them. Neither have they wings to escape by flying, and generally go in large bodies easily found and approached. The use is, I doubt not, the protection the disguise affords against small insectivorous birds. I have found the crops of some humming-birds full of small, soft-bodied spiders, and many other birds feed on them. Stinging-ants, like bees and wasps, are closely resembled by a host of other insects; indeed, whenever I found any insect provided with special means of defence I looked for imitative forms, and was never disappointed in finding them.”
The ant-like species are probably protected by their appearance from the attacks of many of the larger spiders. We have kept great numbers of Attidæ in captivity, and, although they devoured flies, gnats, larvæ, and other spiders, they would never touch ants. Among spiders, however, as among birds, we find that certain groups subsist almost entirely upon ants.
The class of spiders whose mimicry protects them from their enemies, whether they are birds or other spiders, probably includes at least two of our own ant-like species, Synageles picata and Synemosyna formica, which, in confinement, are always hungry for gnats, but will not touch ants, even of small size.
The existence of a class of spiders which mimic the particular species of ants upon which they prey is not to be questioned, but it is doubtful whether the benefit to the spider is increased facility in capturing the ant, or whether it is merely protective. It may be that the spider, by virtue of its resemblance to the ant, not only gets an abundant supply of food, but also escapes being eaten itself, and thus enjoys a double advantage. Both Bates and Wallace take the ground that the advantage derived by the spider consists in greater ease in the capture of prey, but both of these writers refer to spiders only incidentally to illustrate a general proposition, without special consideration of their peculiar conditions.
Mr. Herbert Smith, who has paid a good deal of attention to this subject, is inclined to believe that the mimicry in question is entirely protective. He writes as follows:—
“In the United States there are a few rare spiders that mimic ants. Here at Taperinha we find a good score of species of these spiders aping the various kinds of ants very closely; even the odd, spiny wood-ant, cryptocerus, furnishes a pattern, and there are spiders that mimic the wingless ichneumons. We find, after a while, that the spiders prey upon ants just as our spiders catch flies; indeed, this fact has already been noted by other observers. But we go a step beyond the books when we discover not only that the spiders eat the ants, but that they eat the particular ants which they mimic. At all events, we verify this fact in a great number of cases, and we never find the spiders eating any but the mimicked species.... I do not like to hazard a theory on this case of mimicry. It is difficult to suppose that the quick-witted ants would be deceived even by so close a resemblance; and, in any case, it would seem that the spiders do not require such a disguise in order to capture slow-moving ants. Most birds will not eat ants; it seems likely, therefore, that this is simply another example of protection; the spider deceives its enemies, not its prey; it mimics the particular species that it feeds on, because it is seen in that company when it is hunting, and among a host of similar forms is likely to pass unnoticed.”
At first sight, and especially in view of the fact that such cases are not uncommon among insects, it would be naturally supposed that the object of the mimicry was to enable the spider to approach its victim without exciting suspicion; and it is difficult to account, on any other supposition, for the very close resemblance between certain species of spiders and the particular species of ants which they prey upon. It seems as though the highest point of protective benefit would have been reached long before the resemblance of the spider to the ant had become so close as it really is. On the other hand, it is difficult to believe that ants are deceived, even by those spiders which mimic them most closely, when we remember that their perceptions are so keen that they discriminate not only between ants of their own and different species, but even between ants of their own species living in two different communities.
The mimicry of ichneumon flies by spiders was noted some years ago by Mr. Herbert Smith. This case comes under Class 3, in which one species mimics another which preys upon it. Great destruction is caused by ichneumons which lay their eggs on the bodies of the live spiders, and the disguise probably protects the spider by leading the fly to mistake it for one of its own species.
We have no proof that spiders ever mimic ants as a method of escaping from them, but it is possible that this sometimes happens. We know that some ants prey upon them. The foraging ants of South America destroy spiders as well as many kinds of insects, and Wallace mentions a small, wood-boring ant which fills its nest with small spiders.
If the spiders that feed upon ants deceive them by their mimicry those which are preyed upon by ants would gain an advantage by a similar disguise. I once placed a little ant-like spider of the genus Herpyllus in a bottle with three ants no larger than itself, which I had caught with it in the sweep-net. In a very few minutes the ants had killed and begun to devour the spider. It may be that the resemblance was sufficiently close to deceive them in the open, but failed when spider and ants were confined together in close quarters.