J. HYM. RES.
Vol. 16(2), 2007, pp. 327–336
Natural History and Larval Behavior of the parasitoid Zatypota petronae
(Hymenoptera: Ichneumonidae)
JU-LIN WENG* AND GILBERT BARRANTES1
Escuela de Biologı́a, Universidad de Costa Rica, Ciudad Universitaria Rodrigo Facio, San José,
Costa Rica
__________________________________________________________________________________________________________________________________________________________
Abstract.—The koinobiont ectoparasitoid Zatypota petronae Gauld (Ichneumonidae) parasitizes
medium-sized immatures of the cobweb spider Theridion evexum Keyserling (Theridiidae). Zatypota
petronae apparently attacks the spider inside its retreat. An egg is glued on the antero-lateral dorsal
section of the spider’s abdomen. First-instar larvae remain partially inside the egg chorion which is
attached to the spider’s abdomen. In later instars, a layer of a brownish material (saddle), to which
the 7th and 8th abdominal segments of the larva adhere ventrally, anchors the larva to the spider. In
the last instar the saddle includes the egg chorion and the shed exoskeletons of previous instars. A
row of retractile, dorsal protuberances, crowned with hooklets, is present on abdominal segments 1
to 8 of the final-instar larva. The larva uses the hooklets to grab silk lines of the retreat of the
spider’s web. Hanging on the spider’s web the larva kills the spider and sucks out its body tissues.
Then the larva pushes vigorously laterally with its head against the spider’s corpse, and alternately
presses the corpse against the saddle. These movements, in combination with peristaltic
movements, free the larva from the saddle that falls to the ground with the dead spider. The
larva then constructs its pupal cocoon. Prior to cocoon construction, the larva induces the spider to
reinforce the retreat by adding more threads. Parasitism rate and host behavior are also described.
__________________________________________________________________________________________________________________________________________________________
The polysphinctine pimplines are koino- Parasitoid wasps of the cosmopolitan
biont ectoparasitoids of spiders in several speciose polysphinctine genus Zatypota
families (Nielsen 1923, 1932, Fincke et al. Förster parasitize spiders in at least five
1990, Hanson and Gauld 1995, Gauld et al. families (Dictynidae, Agelenidae, Tetra-
1998). Nielsen (1923, 1932) described in gnathidae, Araneidae and Theridiidae)
detail the behavior of the larvae and hosts (Shaw 1994, Gauld et al. 1998). In the
of several European polysphinctine spe- neotropics the only two host records were
cies. The parasitism rates and life cycle of Theridion species: T. contreras Levi for an
Hymenoepimecis robertsae Gauld on the unidentified Zatypota species (Jiménez
neotropical tetragnathid Nephila clavipes 1987) and T. evexum Keyserling for Z.
(L.) was described by Fincke et al. (1990). petronae Gauld (Barrantes and Weng in
However, larval behavior of neotropical press).
polysphinctine wasps has been described The larval behavior of Zatypota sp.
in detail for only one species, H. argyr- (Jiménez 1987) differs in some aspects from
aphaga Gauld on the tetragnatid Plesiometa that of European polysphinctine species
argyra (Walker) (Eberhard 2000a, 2000b, (Nielsen 1923, 1932) and H. argyraphaga
2001). (Eberhard 2000a). The larva of Zatypota sp.
was said to hold on to the spider by biting
the dorsum or sides of the anterior section
of the spider’s abdomen. This description
* Current address: Department of Entomology, Kansas
State University, Manhattan, Kansas, USA is likely wrong as detailed descriptions of
1 Author for correspondence the behavior of the larva of Z. albicoxa
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:04:53 327
328 JOURNAL OF HYMENOPTERA RESEARCH
(Nielsen 1923) and the larva of H. argyr- lenses. Drawings of larval behavior were
aphaga (Eberhard 2000a) show that some traced from video recordings. Voucher
posterior segments of the larvae lodge specimens of wasps and spiders were
ventrally in a ‘‘saddle’’, probably coagulat- deposited in the Museo de Zoologı́a of
ed spider’s hemolymph, that adheres tight- the Universidad de Costa Rica. Wasp
ly to the spider’s abdomen (Nielsen 1923, species names follow Gauld et al. (1998).
Eberhard 2000a). There is no further in-
formation on the biology of larvae of this RESULTS
Zatypota species. Here we describe the Percentage of parasitism.—Only juveniles
intensity of parasitism and behavior of of T. evexum were found to be parasitized
the larva of Z. petronae and its host T. by Z. petronae. A second instar larva was
evexum. We describe for the first time how feeding on a juvenile spider, possibly
a polysphinctine larva frees itself from the a third instar. However, final instar larvae
spider’s corpse. were found feeding on large immature
spiders, possibly juveniles of fourth to fifth
MATERIALS AND METHODS instars. The parasitism in T. evexum was
Field observations were made from very low (mean percentage of parasitism/
October 2005 to October 2006 in a 250 m2 census 5 1.39%, SD 5 1.80, n 5 53 bi-
plot in the understory of a middle-eleva- weekly censuses). The reproduction in T.
tion wet forest patch (9u 549N, 84u 039W; evexum is extremely seasonal, and the
elevation 1200 m), the Reserva Biológica abundance of immature spiders (4th instar
Leonel Oviedo on the Universidad de or larger) susceptible to attack by Z.
Costa Rica campus, San José Province, petronae increased in March and declined
Costa Rica. All spiders (or nearly so) from drastically through August (Fig. 1). Be-
third-forth instar outside the egg sac to tween September and February the popu-
adults were checked for parasites every lation consists, first, of mature females, and
two weeks; most spiderlings disperse from then of very small spiderlings (Fig. 1).
the mother’s web at fourth instar. The Parasitized spiders occurred primarily
small eggs and early instar larvae probably from March through August.
went undetected. Spider web and wasp attack.—The web of
Theridion evexum constructs most webs T. evexum includes a folded leaf that forms
between 0.20 to 1.5 m above the ground a conical retreat, with a tangle in front of
(Barrantes and Weng in press), making it the retreat opening, and long viscid
possible to find practically all webs. Addi- threads extending from the tangle to other
tionally, we collected seven parasitized leaves (Barrantes and Weng in press). An
spiders and kept them on their webs additional tangle is constructed by the
indoors to observe the behavior of larvae spider inside on the upper side of the
and spiders. In two cases we transplanted retreat.
the plant on which the parasitized spider We witnessed one attack by a female Z.
had constructed its web indoors, allowing petronae wasp. The wasp approached the
us to observe the larva and host behavior web and hovered in front of the spider’s
with little disturbance. The complete larval retreat opening. The wasp then flew inside
development was not observed in all cases, the retreat. A few seconds later, the spider,
so sample sizes are not always the same. with the wasp perched on its dorsum,
Behavior and morphological features of dropped about 10 cm below the retreat,
more than 10 larvae were observed under and hung on its dragline. They struggled
a dissecting microscope. Video recordings for a few seconds and then the wasp flew
of behavior were made using a Sony DCR – out of sight. The spider began to climb
VX 1000 camcorder with +5 close-up towards the retreat but after advancing
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:05 328
VOLUME 16, NUMBER 2, 2007 329
Fig. 1. Temporal changes in the number of immature spiders (black rhombus), males (open circles), and
reproductive females (black circles) of T. evexum.
about four centimeters, it became para- ‘‘second’’ instar, larvae (n 5 5) were
lyzed and fell back, motionless for about completely outside the collapsed, flattened
10 min, dangling from its dragline. The egg chorion that was embedded in an
spider recovered its motion slowly, and apparently rigid, semitransparent layer of
with clumsy movements cleaned some of brownish material (Fig. 2B) (the ‘‘saddle’’
its legs before ascending to the retreat. of Nielsen 1923). The ventral surface of two
When we returned, 30 min later, the spider or three posterior abdominal segments
had fully recovered its mobility. We did rested on the saddle. In subsequent instars,
not ascertain whether the spider had an the cuticles of the previous molts became
egg on its abdomen. incorporated into the saddle as they ad-
Larvae.—We observed one egg of Z. hered to its upper surface, against the
petronae glued on the antero-lateral dorsal ventral surface of the larva. The egg
section of the abdomen of a spider collect- chorion was near the spider’s surface, but
ed in the field, a first instar larva emerged not in contact with it. The saddle was
about four hours later. All eleven larvae of attached by a short pedicel to the spider’s
different instars checked under the dissect- abdomen (Fig. 2B), and the larva’s abdom-
ing microscope were attached by their rear inal segments 7 and 8 secured it to the
end to the cuticle of the antero-lateral saddle. Feeding scars were observed on the
surface of the spider’s abdomen (Fig. 2A). nearby dorsal and lateral surface of the
The first instar larva (n 5 3) had its spider’s abdomen (Fig. 2B).
posterior end lodged inside the egg chori- In the final instar, larvae had dorsal,
on, with its head, thorax, and some two-lobed, retractable tubercles on eight
abdominal segments protruding; the cho- abdominal segments (1st to 8th); these
rion remained attached to the spider. In the structures were absent in previous stages.
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:05 329
330 JOURNAL OF HYMENOPTERA RESEARCH
Fig. 2. Larva of Z. petronae: A- Second instar larva attached to the antero-lateral surface of the spider’s
abdomen. B- Penultimate larva with the saddle attached to the spider’s abdomen. The shed cuticles of previous
molts are visible under the larva. Feeding scars (black dots) are also visible on the surface of the spider’s
abdomen. (Photo of a specimen in alcohol).
When extended, the tubercles were movement. The larva fed first on the
crowned with a circle of tiny hooks that spider’s abdomen, then on its cephalotho-
allowed the larva to grab the threads of the rax. When discarded, the spider’s carcass
spider’s web inside the retreat. The larva was nearly completely empty; even its legs
could extend or retract independently each were almost transparent. The larva was
lobe of the tubercle, and the tubercles could thus capable of extracting nearly complete-
be retracted rapidly and completely into ly the spider’s internal tissues, presumably
a pocket. Based on size and morphology, using capillarity (Eberhard et al. 2006).
we discerned three instars in the larvae of Dislodging the saddle.—After the larva
this wasp. However, the saddle of what we had finished feeding, it began to free itself
thought was a second instar larva included from the saddle while hanging inside the
the chorion and the shed cuticles of two spider’s retreat. The process, which lasted
molts. Hence, further observations are about two hours, included three types of
needed to confirm the number of instars. movement: pressing the spider carcass
The final instar larva spent about 18 h against the saddle, pushing the spider
attached to the spider (n 5 2), three to six carcass laterally, and peristaltic move-
hours after removing the saddle and prior ments of the larva’s abdominal segments.
to cocoon construction (n 5 5), and nearly The pressing and peristaltic movements
18 h constructing the cocoon (n 5 1). The seemed to be more frequent and intense
duration of the larva inside the cocoon just before the spider carcass and saddle
before pupation was not recorded. One were completely removed.
penultimate instar larva molted during the Pressing movements: The ventral side
night and the next morning hung from of the larva’s head pushed on the spider’s
lines near the roof of the retreat with its anterior end, steadily pressing the spider’s
dorsal hooks, and fed on the spider for carcass against the saddle until it bent
about eight hours. During approximately almost completely over the saddle
the first four hours the spider’s legs moved (Fig. 3A, B). The larva then released the
slightly, but later we could not detect any pressure completely as it moved its head to
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:06 330
VOLUME 16, NUMBER 2, 2007 331
Fig. 3. Movements of the spider to free itself from the saddle (traced from video images). Pressing movement:
the larva places its head near the spider’s chelicerae (A) and presses the spider carcass against the saddle (B).
Lateral pushing: the larva places its head on the anterior tip of the spider carcass and pushes it laterally (C). A
backward final push completely dislodges the saddle from the larva (D); the grey arrow shows the position of
the larva’s head tip before pushing the saddle. Dotted and dashed lines represent the initial positions of the
larva and the spider’s carcass respectively.
the initial position, and then either made oriented at about 30u to its longitudinal
another pressing movement or pushed the axis and contacted the saddle, rather than
spider’s carcass laterally (see below). the spider and the force exerted by the
Lateral pushing: The larva bent ventral- pushing movement was toward the rear of
ly until the lateral section of its head the larva’s body rather than laterally
contacted the legs and/or cephalothorax (Fig. 3D).
of the dead spider, and then pushed Peristaltic movements: Peristaltic waves
laterally (Fig. 3C). Then it moved its head moved posteriorly along the larva’s abdom-
back slightly, maintaining contact with the inal segments during pressing and pushing
carcass, and pushed laterally again. The movements. The last segment stretched
larva pushed repeatedly up to 10 times extensively backward as the wave reached
before reorienting its head; the complete it. The peristaltic waves were strongest
carcass moved visibly with each push by during the last pressing and lateral move-
the larva. The larva often placed its head ments of the larva.
on the opposite side of the spider during Final events: As soon as the saddle was
successive pushing bouts. During the last released the larva rubbed its head against
three pushing bouts the larva’s head was the ventral surface of the segments that
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:10 331
332 JOURNAL OF HYMENOPTERA RESEARCH
had been connected to the saddle, which All 57 cocoons found were constructed
were covered with a mucilaginous sub- inside the spiders’ retreats, but their
stance. The small processes on the ventral attachment varied among retreats: 71%
larval segments that are inserted in the were attached to the threads of the tangle
saddle in other polysphinctines (the ‘‘taps’’ near the retreat’s roof, 20% were attached
of Nielsen 1923, Eberhard 2000a) were not to the threads applied by the spider at the
visible in the Z. petronae larva at the apex of the leaf-cone (Fig. 5A), and 9%
moment the saddle was released. Howev- were in the middle of the retreat, attached
er, examination of two saddles under the to a thick silk cable formed by several
dissecting and compound microscopes independent threads (Fig. 5B).
showed a wedge-like depression inside Enemies of the wasp.—Of the 57 cocoons
the saddle. This depression was likely found, we observed two predation attacks
produced by an abdominal projection that and a possible parasitoid attack on a third
anchored the larva to the saddle. cocoon. One pupa was attacked by Sole-
Cocoon construction.—One larva of Z. nopsis ants inside the spider’s retreat. A
petronae began cocoon construction at second pupa or larva inside its cocoon was
about 18:30 h inside the spider’s retreat, attacked by a penultimate male of T.
after resting for nearly two hours. We did evexum that fed on the immature wasp
not follow cocoon construction in detail, through the cocoon silk. The third cocoon
but our incomplete observations indicate had a lateral hole near its bottom that
that the behavior was quite similar to suggested the exit of a parasitoid, as adults
cocoon construction by H. argyraphaga of Z. petronae exit the cocoon by cutting
(Eberhard 2000a), except that no suspen- a circular slit near the cocoon’s upper end.
sion line was built. Construction lasted Host spider behaviour.—The spiders car-
nearly 18 h (N51). It began with the larva rying first and possibly young second
hooked by its dorsal tubercles to the silk instar larvae (N54) were capable of cap-
threads of the tangle inside the retreat turing prey trapped on the long viscid lines
(Fig. 4A). of their webs. Their attack behavior was
The larva built the cocoon by attaching indistinguishable from the attacks of non-
a silk line (or lines) produced from its head parasitized spiders (Barrantes and Eber-
to the tangle of threads made by the spider, hard in prep.). However, spiders with
and pulling its head from this point to the a large penultimate instar or a final instar
next attaching point, which was either larva did not attack prey that adhered to
another tangle thread or one of its own the sticky threads. The stickiness of the last
previously produced lines. Cocoon con- capture threads produced by a spider with
struction began around the posterior por- a large penultimate instar larva was nota-
tion of the larva (Fig. 4B) and then gradu- bly reduced, as Drosophila flies (with their
ally extended upward until it enclosed the wings cut) walked easily along these
larva. The first silk lines around the larva threads.
formed a loose, fluffy mass (Fig. 4C), but On four occasions we observed that
after some hours a much denser wall began when a larva apparently bit the cuticle of
to form around the larva (Fig. 4D). The a spider’s abdomen, the spider jerked and
larva frequently paused during the con- tried unsuccessfully to reach the larva with
struction for up to 2 min. After 20 h the its legs I, II and III. This suggests that the
larva ejected its meconium through the spider perceived and was irritated by the
circular hole at the bottom of the cocoon. wounds produced by the larva. In one case
The recently constructed cocoon had a pale- the spider’s leg II touched the anterior
yellow color that turned to orange-yellow portion of the larva, and the larva imme-
over the next day. diately moved its anterior portion toward
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:12 332
VOLUME 16, NUMBER 2, 2007 333
Fig. 4. Sequence of cocoon construction. A) Final instar larva recently freed from the saddle, dorsal tubercles
are visible on two abdominal segments. B) Larva about 45 min after cocoon construction began. C) Cocoon
construction after approximately 2 h. D) Cocoon after 20 h; note the meconium below the cocoon.
the dorsal-middle section of the spider’s across the retreat opening (72%, n 5 57;
abdomen (out of range of the leg) and Fig. 5B), inside, more or less in the middle
apparently bit her again. Examination with of the retreat (20%) (Fig. 5B), or both (8%)
a hand lens showed that there was a tiny across the retreat opening and inside it
shiny spot, presumably of hemolymph, (Fig. 5B). In one case the threads inside the
where the larva had apparently first retreat were so dense that they formed
bitten the spider (documenting that the a sketchy sheet just below a cocoon
larva actually bit the host rather than (Fig. 5C), which was attached to the tangle
just touched it with its mouthparts is not threads. A parasitized spider added more
easy). threads to the apex of the retreat (Fig. 5A),
The web retreats housing cocoons had possibly during the last two nights, before
additional, non-sticky thick threads either being killed by the larva.
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:12 333
334 JOURNAL OF HYMENOPTERA RESEARCH
Fig. 5. Retreat constructed by Theridion evexum. A) The arrow shows the threads that maintain the leaf-retreat
folded. B) Threads added by the spider at the retreat opening (a) and in the middle of the retreat (b). C) Sketchy
sheet in the middle of the retreat. Larva of Zatypota petronae induces T. evexum to produce threads at the retreat
opening, inside it and to increase number of threads that maintained the leaf folded (black arrow in A).
DISCUSSION 2001), whereas the larva of Reclinervellus
The morphology and behavior of the nielseni (Roman) [5 Polysphincta nielseni]
larva of Z. petronae are quite similar to (Nielsen 1923, Gauld and Dubois 2006) and
those of larvae of other polysphinctine P. gutfreundi Gauld (Gauld et al. 1998),
species (Nielsen 1923, 1932, Fincke 1990, which also lack suspension lines, attach the
Gauld et al. 1998, Eberhard 2000a). How- cocoons to the threads near, or on the hub
ever, they often differ in where and of the orbicular web of Cyclosa conica
possibly how their cocoons are attached (Pallas) (Nielsen 1923) and Allocyclosa
to the host web. The larva of Z. petronae bifurca (McCook), respectively. These dif-
attaches its cocoon, which lacks a suspen- ferences are likely determined by the
sion line, to silk threads inside the retreat characteristics of the web of each spider
of T. evexum, Hymenoepimecis spp. attach species, particularly by the modifications
their cocoon to the spider web (e.g. N. of the web (the ‘‘cocoon web’’ of Eberhard
clavipes and P. argyra) with a suspension 2001) induced by the parasitoids (e.g. T.
line (Fincke et al. 1990, Eberhard 2000a, evexum re-enforcing its retreat).
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:14 334
VOLUME 16, NUMBER 2, 2007 335
There are also differences in how larvae cable of silk threads produced inside the
adhere to the saddle. Larvae of Z. petronae retreat was not found in webs of un-
apparently adhere to the saddle using parasitized spiders. The reinforcement of
wedge-like projections of one or two seg- the retreat with additional silk threads
ments, rather than taps as in Z. albicoxa and possibly increases the protection of the
H. argyraphaga. Differences may also exist in cocoon, primarily against heavy rains,
the sensitivity of the host to the wounds which is likely important for the wasp’s
caused by the parasitoid. For example, P. survival. If a retreat opens up, it is unlikely
argyra did not show any reactions to that the thin threads of the tangle inside the
apparent bites of H. argyraphaga larvae retreat, where most cocoons were attached,
(Eberhard 2000a). However, T. evexum could survive heavy rains intact.
reacted by jerking its body and moving its Our observations suggest that Z. petronae
legs toward the point where the larva was is not specialized on a particular species of
biting the spider’s cuticle. This suggests that host. This wasp parasitized intermediate
chemical composition of secretions could sized spiders (at least 4th instar), but the
vary among parasitoid species. Further reproduction of T. evexum is highly sea-
research to confirm chemical differences in sonal and large juvenile spiders occur only
the saliva of parasitoids and differences in during five or six months of the year. Thus,
sensitivity of spider hosts to the bites of it is likely that Z. petronae must parasitize at
their parasitoids is needed. least one other species of spider to main-
The release of the saddle by final instar tain its population.
larvae is much more complex than simply The percentage of parasitism of T.
the muscular movements of the posterior evexum (1.39% 6 1.80) was relatively low
end of the larva as suggested by Nielsen when compared with other spider species.
(1923) and Eberhard (2000). Without the Fincke et al. (1990) reported that the annual
powerful pressing and pushing move- percentage of parasitism for intermediate-
ments of the larva against the saddle, the sized juvenile females of N. clavipes was
peristaltic abdominal movements are pos- 15–30%, and Eberhard (2000) reported that
sibly insufficient to free it from the saddle. the parasitism on P. argyra was higher than
More information is needed to examine the 40% for mature females and higher than
possible differences among polysphinctine 3% for mature males. The low parasitism
species. on T. evexum also suggests that Z. petronae
The larva of Z. petronae induces the host is possibly using other spiders as hosts in
spider to add more threads on different the same area.
sections of the retreat (apex, inside, and
across the retreat opening) that make this ACKNOWLEDGEMENTS
structure stronger and more durable. Add-
We thank Ingi Agnarsson and Paul Hanson for
ing threads near the apex of the retreat is identifying the spider and wasp respectively, and
apparently a repetition of a subroutine William G. Eberhard, Paul Hanson, Mark Shaw, and
used in the construction of the retreat by an anonymous reviewer for valuable comments on
an unparasitized spider, since threads previous drafts. We also thank Andrea Bernecker for
her comments that greatly improved the drawings.
applied in similar fashion allows the spider
to fold the leaf and maintain the retreat’s LITERATURE CITED
shape. Similarly, threads across the retreat
opening were occasionally present (3 out of Agnarsson, I. 2004. Morphological phylogeny of
17 webs) when pre-adult female spiders cobweb spiders and their relatives (Araneae,
Araneoidea, Theridiidae). Zoological Journal of the
were molting, though these threads were Linnean Society 141: 447–626.
not as abundant as those in retreats of Barrantes, G. and J. L. Weng. In press. Natural history,
parasitized spiders. However, the thick courtship, feeding behaviour and parasites of
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:15 335
336 JOURNAL OF HYMENOPTERA RESEARCH
Theridion evexum (Araneae: Theridiidae). Bulletin Gauld, I. D. and J. Dubois. 2006. Phylogeny of the
of the British Arachnological Society. Polysphincta group of genera (Hymenoptera:
Eberhard, W. G. 2000a. The natural history and Ichneumonidae; Pimplinae): a taxonomic revision
behavior of Hymenoepimecis argyraphaga (Hyme- of spider ectoparasitoids. Systematic Entomology
noptera: Ichneumonidae) a parasitoid of Plesio- 31: 529–564.
meta argyra (Araneae: Tetragnathidae). Journal of ———, J. A. Ugalde G., and P. Hanson. 1998. Guı́a de
Hymenoptera Research 9: 220–240. los Pimplinae de Costa Rica (Hymenoptera:
———. 2000b. Spider web manipulation by a wasp Ichneumonidae). Revista Biologı́a Tropical 46
larva. Nature 406: 255–256. (Supl. 1): 1–189.
———. 2001. Under the influence: webs and building Hanson, P. and I. D. Gauld. 1995. The Hymenoptera of
behaviour of Plesiometa argyra (Araneae, Tetra- Costa Rica. Oxford University Press, Oxford.
gnathidae) when parasitized by Hymenoepimecis Jiménez, M. L. 1987. Relaciones entre arañas y avispas.
argyrophaga (Hymenoptera, Ichneumonidae). Folia Entomológica Mexicana 73: 173–183.
Journal of Arachnology 29: 354–366. Nielsen, E. 1923. Contributions to the life history of
———, G. Barrantes, and J. L. Weng. 2006. The the pimpline spider parasites (Polysphincta, Za-
mystery of how spiders extract food without glyptus, Tromatobia). Entomologiske Meddelelser 14:
masticating prey. Bulletin of the British Arachnolo- 137–205.
gical Society 13: 372–376. ———. 1932. The Biology of Spiders. Levin & Munks-
Fincke, O. M., L. Higgins, and E. Rojas. 1990. gaard, Copenhagen.
Parasitsm of Nephila clavipes (Araneae: Tetra- Shaw, M. R. 1994. Parasitoid host ranges. Pp. 111–144
gnathidae) by an ichneumonid (Hymenoptera, in: B. A. Hawkins, and W. Sheehan, eds. Parasitoid
Polysphinctini) in Panama. Journal of Arachnology community ecology. Oxford University Press, Ox-
18: 321–329. ford.
Journal of Hymenoptera Research hyre-16-02-09.3d 25/7/07 18:05:15 336