ARG YRODES A TTENUA TUS (THERIDIIDAE): A WEB THAT IS NOT A SNARE*

Spiders of the large theridiid genus Argyrodes, whose natural history was reviewed by Exline and Levi (1962), seem to have generally abandoned the usual theridiid habit of spinning webs to capture insect prey. A few spin their own webs, but more often they live in the webs of other, larger web-building spiders where they remove prey from the host’s web (e.g. Kullmann 1959, Vollrath 1978). The apparently kleptoparasitic species A. trigonum .(trigonum species group) actually preys on its hosts on occasion. Argyrodes species in the Rhomphaea group have also been found feeding on web spiders, but some in circumstances which suggest a strict predator-prey relationship rather than kleptoparasitism. Exline and Levi (1962) note two observations of A. (R.)fictilium preying on web weavers although fictilium "often make small theridiid webs of their own". I have found an A. (R.)projiciens (O.P. Camb.) feeding on a Metazygia sp. which leaves its web up only during the night, thus making kleptoparasitism unlikely; the state of the prey’s web indicated that it was attacked as it was building its web early in the evening. I have also seen an unidentified species feeding on an Araneus (?) sp., a species which leaves its web up only two to three hours in the early evening and then removes it completely. Clyne (1979) showed that a species from the Ariamnes group, Argyrodes colubrinus, uses still another tactic. These spiders also specialize on spider prey, but attack ballooning mature male spiders which walk onto their webs of long, non-sticky threads. Roberts (1952) notes that A. (A.)flagellum also spins a web consisting of only a few long, very fine theads; perhaps mistakenly he states that the lines are sticky.


INTRODUCTION
Spiders of the large theridiid genus Argyrodes, whose natural history was reviewed by Exline and Levi (1962), seem to have generally abandoned the usual theridiid habit of spinning webs to capture insect prey. A few spin their own webs, but more often they live in the webs of other, larger web-building spiders where they remove prey from the host's web (e.g. Kullmann 1959, Vollrath 1978. The apparently kleptoparasitic species A. trigonum .(trigonum species group) actually preys on its hosts on occasion. Argyrodes species in the Rhomphaea group have also been found feeding on web spiders, but some in circumstances which suggest a strict predator-prey relationship rather than kleptoparasitism.
Exline and Levi (1962) note two observations of A. (R.)fictilium preying on web weavers although fictilium "often make small theridiid webs of their own". I have found an A. (R.)projiciens (O.P. Camb.) feeding on a Metazygia sp. which leaves its web up only during the night, thus making kleptoparasitism unlikely; the state of the prey's web indicated that it was attacked as it was building its web early in the evening. I have also seen an unidentified species feeding on an Araneus (?) sp., a species which leaves its web up only two to three hours in the early evening and then removes it completely. Clyne (1979) showed that a species from the Ariamnes group, Argyrodes colubrinus, uses still another tactic. These spiders also specialize on spider prey, but attack ballooning mature male spiders which walk onto their webs of long, non-sticky threads. Roberts (1952)  This note describes the web and predatory behavior of another species of the Ariamnes group, Argyrodes attenuatus, which attacks ballooning immature and mature male spiders, and also minute "trapeze" flies which use its web as a resting place. This spider's web serves not as a trap, but rather as a resting site for both types of prey, and as a substrate for the spider's stealthy attacks.

OBSERVATIONS
Argyrodes attenuatus is widely distributed in Central and South America (Exline and Levi 1962). The observations reported here were made on Hacienda Mozambique about 15 km SW of Puerto Lopez, Meta, Colombia (el. approx. 200 m) between June and August 1978 in two patches of periodically flooded forest from which some trees had been cut. The forest at one site was in at least an advanced secondary stage with some relatively large trees, while the other was younger. Voucher specimens have been placed in the Museum of Comparative Zoology, Cambridge, Mass.

Resting Position and Web
More than 40 spiders were seen from 1-3 m above the ground, resting on sparse, irregular, three-dimensional networks of long threads (Fig. 1). The networks ranged from two or three up to seven threads, some of which reached lengths of up to 1-2 m; careful inspection showed that none of the threads were sticky.
During the day the spiders usually rested with all their legs pressed tightly together, with I and II directed forward and III and IV backward, and the abdomen straight or nearly so, giving the animal a stick-like appearance. At night the legs, especially the front ones, were often spread. The spider always rested at a junction of two or more threads. Although the spider could control the form of its long thin abdomen, and sometimes curled it into a tight spiral, I never saw an individual move it in an undulatory or inchworm-like fashion as G61di (in Exline and Levi 1962) saw A. (A.) longissimus do I did not disturb the spiders as G61di apparently did, however). The adaptive significance of the extraordinary abdomen of A. attenuatus may be at least partly outline camouflage rather than imitation of inch worms as proposed by Exline and Levi (1962).
The spiders replaced the threads they walked along as they moved, and thus probably held one broken end with a leg and spanned the gap with the body as they rested on the web, but they did not tense the lines or sag them suddenly when disturbed as do other spiders with similar reduced webs such as Miagrammopes (Akerman 1932, Lubin et al. 1978 and Ulesanis (Marples 1955).

Attack Behavior and Prey
Twenty seven spiders were found feeding on or resting near enswathed prey (Table 1). In addition I saw mature females move after but fail to capture a mature male theridiid spider and a small unidentified spiderling on lines in their webs. Some of the spiders' prey were judged to be substantially heavier than the A. attenuatus which captured them. In general, small A. attenuatus fed on small flies, and larger ones fed mostly on immature non web-building spiders. The ways these prey were secured was determined by direct observation and experiment.
When small spiderlings of several species were placed on webs of mature female A. attenuatus, the owners consistently responded by moving toward them and attempting to attack. In a typical capture sequence a spiderling was held dangling from its trail line, and  attenuatus advanced toward it slowly and smoothly. When the female was about 15 cm away, the spiderling began walking toward her along the line, and she suddenly began approaching more rapidly. Then, before she had reached the advancing spiderling, she quickly turned 180 and pulled a line of wet sticky silk from her spinnerets with her legs IV, apparently contacting the line with the outer surfaces of the tarsi (Figs. 2 and 3). The instant the spiderling came in range, she wrapped it in this thread with a quick movement of the hind legs, then bit it and wrapped it further in apparently dry silk. Finally she carried it dangling from her leg III to the place where she had been resting, and began to feed on it. Aggressive responses such as these were usually elicited when spiderlings were placed on occupied webs, but Philoponella sp. and Spilasma artifex spiderlings consistently escaped because they did not stay on the A. attenuatus lines, but descended on lines of their own, and as the A. attenuatus approached them or tried to reel them in, they let out silk faster until the A. attenuatus finally turned away.
As noted by Eberhard (1980) and Lahmann and Zufiiga (in prep.), it is common to find small insects, mostly nematocerous flies, hanging at night from spider threads in some forested areas. These "trapeze" flies seem to prefer more or less horizontal lines, but do not discriminate between threads from different spider species, and will even hang on cotton thread (Lahmann and Zufiiga in prep.). Young A. attenuatus were frequently found with small flies as prey, and the supposition that these prey had been captured while hanging on the spiders' webs was reinforced when I observed an attack on such a fly. When I first noticed the fly, it was hanging on one of the threads of a spider's web, and the spider was about 5 cm away, facing toward it. As I watched, the spider moved very slowly and smoothly toward the fly, stopping frequently to hang motionless; it took nearly five minutes to cover the 5 cm. There were occasional gentle vibrations of the line, but I could not tell whether they were caused by the spider or a sporadic gentle breeze. As the spider neared the fly, it began waving its front legs slowly; it touched the fly Psyche [December gently with one of these movements, and immediately withdrew its leg and turned slowly 180 so that its abdomen was toward the prey. Carefully it drew out a strand of sticky silk with first one IV and then the other (as in Figs. 2 and 3) and then suddenly, with the only quick movement of the entire sequence, swung its hind legs. upward to bring the sticky thread around the fly, and immediately began wrapping it with dry silk.

Egg Sac
Five different A. attenuatus were found with egg sacs. All sacs were elongate and white with a soft papery wall and small yellowish eggs loose inside the central portion. The spider rested either under the sac or adjacent to it with the legs extended in front and behind so that its outline merged with that of the sac.

DISCUSSION
One of the unusual features of A. attenuatus is the presence of serrated bristles on the prolateral side of tarsus IV, and the absence of the comb on its inner surface that is usual for theridiids (Exline and Levi 1962). It is thus interesting that, at least in the si'ngle case in which leg and thread positions were observed in detail (right leg IV in Fig. 3), the sticky wrapping thread contacted the outer rather than the inner surface of the leg. This association is in accord with the general belief that serrated bristles and combs on the tarsi IV of theridiids serve to manipulate the wet wrapping silk which is apparently peculiar to this family. The advantage of holding the thread on the outer rather than the inner surface of the tarsus is not clear; perhaps it is related to the way in which the spider holds sticky thread poised just before wrapping (Fig. 2), a behavior unique to my knowledge in theridiids.
The web structure, attack behavior, egg sac, and cryptic postures of A. attenuatus appear to be very similar to those of A. colubrinus (Clyne 1979). A. flagellum also apparently assumes the same cryptic posture. A. attenuatus differs from what is known of the other two in that it captures small flies and immature spiders as well as mature male spiders, apparently applies less wet silk as it wraps prey, and does not always rest facing up on webs. These spiders appear to be unique in having webs that serve as attractive resting sites and walkways for their prey.