Content uploaded by M. E. A. Whitehouse
Author content
All content in this area was uploaded by M. E. A. Whitehouse on Nov 26, 2017
Content may be subject to copyright.
J.
Zool.,
Lond
(A)
(1986) 210,279-303
The biology
of
New Zealand and Queensland pirate spiders (Araneae, Mimetidae):
aggressive mimicry, araneophagy and prey specialization
R.
R.
JACKSON
AND
MARY
E.
A.
WHITEHOUSE
Depurtment
of
Zoology, University
of
Canterbury, Christchurch
I,
New Zealand
(Accepted
19
November
1985)
(With
1
plate and
4
figures
in
the
text)
Mimelus
sp. indet. and
Mimetus
maculosus,
from New Zealand and Australia, respectively, were
studied in the laboratory and in nature. Behaviourally, the two species were very similar. Each
was found to be primarily an araneophagic spider which invaded alien webs, acted as an
aggressive mimic by performing a variety of vibratory behaviours to which the prey-spider
responded as
it
normally would to its own prey, and attacked by lunging at close range, subduing
its victim with a strong, apparently spider-specific venom while holding the spider in a 'basket'
formed by its spine-covered legs. In nature, these mimetids were observed
to
feed
on
a restricted
range of spiders: orb web-building araneids and space web-building theridiids. Sometimes, they
occupied other types of webs, but in the laboratory they captured only araneids and theridiids
efficiently. They captured non-cribellate amaurobiids considerably
less
efficiently, and never
captured other types of spiders. Occasionally, the mimetids fed on insects ensnared in araneid
and theridiid webs and on eggs of theridiids. Experimental evidence indicated that vision was
of little
or
no importance in the predatory behaviour of these mimetids. The behaviour
of
the
mimetids is compared to that of
Porriu,
an araneophagic web-invading salticid, and the results
of this study are discussed in relation to hypotheses concerning salticid evolution.
Contents
Introduction
........................
Materials and methods
..................
Testing
..........................
General
........................
Type
B
and informal tests..
Type
A
tests
......................
..............
Pursuit tendency, capture rate and capture efficiency
....
Tests in light-free chamber and tests with painted spiders
Observations
........................
Observations in nature
..................
Morphology and appearance
..............
Locomotion
......................
Elements of predatory behaviour
............
Spread legs posture
..................
Cocked legs posture
..................
Up-and-down plucking
................
Pedal plucking
....................
Tugging
........................
Pulsating
........................
Tapping
........................
Lunging ........................
279
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
............
Page
280
28
I
28
1
28
I
283
284
284
284
285
285
285
286
286
286
286
287
288
289
290
290
290
0022-S460/X6/0
10279
+
25
803.00
@
1986
The Zoological Society
of
London
280
R.
R.
JACKSON AND
M.
E.
A.
WHITEHOUSE
Post-attack behaviour
....................
Wrapping
........................
Carrying prey.
.......................
Drop
on
prey
........................
Sequences
of
behaviour while
on
araneid and theridiid webs
.
.
Predation
on
spiders
....................
Predation
on
insects
....................
Oophagy
..........................
Multiple prey-captures
..................
Comparison of
Mimetus
sp. and
Mimetus maculosus
......
Sequences of behaviour while
on
other web-types
Sparsely woven non-sticky webs
..............
Densely woven and cribellate webs
Pursuit tendency and capture efficiency
Observations
on
other mimetid species
......
............
............
Tests in light-free chamber and with painted spiders
......
Discussion
..........................
Web invasion, araneophagy and aggressive mimicry
......
Kleptoparasitism and oophagy
................
Locomotion
........................
............
Sensory control
of
predatory behaviour
............
Capture efficiency and pursuit tendency
............
The relationship between predatory behaviour and acute vision
The evolution of the Mimetidae
................
References
..........................
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
29
1
29
I
293
293
293
293
293
294
295
295
295
295
296
296
296
291
298
298
299
299
299
299
300
301
302
Introduction
Although spiders are often referred to as either web-builders or cursorial hunting spiders,
certain species cannot be assigned
so
simply to one or the other of these groups. Web-invaders,
spiders which routinely use the webs of other spiders, form one
of
the most important
of
the
aberrant groups.
The mimetids, sometimes called ‘pirate spiders’ (Bristowe,
1958),
are widely regarded as a
family of web-invading araneophagic spiders which use vibratory aggressive mimicry (e.g. Curio,
1976;
Foelix,
1982).
However, there are few published observations on species from this large
cosmopolitan family, and no detailed studies on any species have been carried out. Most of what
is known comes from behavioural observations on
Ero furcata
by Gerhardt
(1924),
Bristowe
(1941)
and Czajka
(1963).
Gerhardt and Bristowe reported that
E.furcata
moves
on
to the edges
of theridiid webs and ‘by gently jerking on the web’, mimics the struggles
of
a trapped insect and
lures the resident spider to within attacking distance. Czajka observed an
E.
furcata
moving on
to a web containing a courting pair of theridiids. The movements
of
the mimetid resulted in the
theridiid male moving away. The mimetid took up the position vacated by the theridiid male,
began moving slowly towards the female while using its forelegs to beat on threads, and made
an attack on, killed, and ate the theridiid female when it got close. Czajka interpreted the
interaction as a case of the mimetid mimicking the courtship of theridiids.
Since the mimetids are widely regarded as major examples of predators that use aggressive
mimicry, an unusual predatory technique (Wickler,
1968),
there is clearly a need for precise and
detailed information concerning the behaviour of additional species from this family. Although
numerous species of mimetids occur in Australia and New Zealand, virtually nothing is known
about the behaviour of any of the Australasian mimetids. Taxonomic ignorance of a group often
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
28
1
discourages behavioural studies, and except for Heimer (In press), virtually no taxonomic work
has been carried out on mimetids of this region.
This paper will be concerned with the behaviour of two species of mimetids,
Mimetus
sp. indet.
from New Zealand (which will be referred to simply as
‘Mimetus
sp.’) and
M.
maculosus
Rainbow
from Australia. Voucher specimens of New Zealand and Australian mimetids have
been deposited at the Otago Museum (Dunedin) and the Queensland Museum (Brisbane),
respectively. This will be the first formal study
of
any Australasian mimetid species.
Mimetid biology is also of interest in relation to recent studies of web-invading spiders from
another family, the Salticidae, which are unique because of their complex eyes and acute vision
(Land, 1969a,
b;
Williams
&
McIntyre, 1980). Salticids are generally considered to
be
the classic
examples of cursorial hunting spiders which, instead of building webs to ensnare their prey, use
vision to stalk, chase and leap on active insects (Drees, 1952; Forster,
L.
M., 1982). However, a
small group of salticid species, all belonging to the primitive subfamily Spartaeinae, are known
to be araneophagic spiders that invade diverse types of webs and practise vibratory aggressive
mimicry: five species of
Portia,
two species of
Brettus
and one species of
Cyrba
(Jackson
&
Blest,
1982; Jackson
&
Hallas, In press
a,
b).
Because of the potential importance of the web-invading
spartaeines for understanding salticid evolution, the salticids and the mimetids will
be
compared
in the Discussion of this paper, and possible implications of these comparisons will be considered
in relation to current speculations about salticid phylogeny.
Materials and methods
Both species were collected and observed in the field.
Mimetus
sp. was found at all times of the year on
fences and the sides
of
buildings
in
Christchurch (South Island, New Zealand).
Mimetus
maculosus
was
found during the early parts of the rainy seasons (November-January) of
3
years (1980-1982) in rain
forests near Cairns in Queensland, Australia (for details concerning the habitat, see Jackson
&
Blest, 1982).
Webs
of
other species were numerous in the habitats
of
both species.
Both species were also studied
in
the laboratory, where they were maintained and tested in transparent
plastic cages and glass tanks
in
rooms with controlled light (12L: 12D) and temperature
(c.
25 “C).
Laboratory studies were carried out
on
large juveniles and adult females of each species
(c.
50
individuals
of each species).
Maintenance and testing procedures, terminology and conventions for describing behaviours (e.g. defini-
tions
of
phase, amplitude, etc.) were as in the study
of
Porria
(Jackson
&
Hallas, In press
a).
Amplitudes
of movements tended to be proportional to the spider’s size. To simplify description, the amplitudes given
in the text
will
be those appropriate for adult females
of
Mimetus
sp.
Observations were carried out during both the ‘day’ and the ‘night’ phases
of
the laboratory photoperiod,
dim red light being used at ‘night’. There was more spontaneous locomotion on the part
of
the mimetids
at night. Otherwise, day- and night-time behaviours were not distinguishable.
Testing
General
The spiders and insects with which the mimetids were tested, and the types of tests for which each
species was used, are indicated in Table
I
(additional information about each species: Jackson
&
Hallas,
In press
a).
Insects were adults; spiders included both juveniles and adults. ‘Small’ prey were 0.25-0.5 times the size
of the mimetids with which they were tested; ‘large’ prey were about equal
in
size to, or slightly larger
282
R.
R.
JACKSON AND M.
E.
A.
WHITEHOUSE
TABLE
1
Spiders and insects used in
tests
with mimetids
Family Species Description Observations
~ypes
or
tests* Size** from field$
Amaurobiidae
Araneidae
Pholcidae
Pisduridae
Theridiidae
Ulo
boridae
Muscidae
Tipulidae
Badumna insignis
Badumna
(L.
Koch)
longinquus
(L.
Koch)
Cam bridgea
antipodiana
(White)
Araneus pustulosu~
(Walckenaer)
Cyclosa anatipes
(Keyserling)
Nephila maculata
(Fabricius)
Pholcus ancoralis
L.
Koch
Pholcus
phalangioides
(Fuesslin)
Psilochorus
sphaeroides
(L.
Koch)
Inola subtilis
Davies
Achaearanea
camura
(Simon)
Achaearanea
mundula
(L.
Koch)
Achaearanea
ventricosa
(Rainbow)
Achaearanea
sp.
indet.
Lutrodectus
katipo
Powell
Philoponella
variabilis
(Keyserling)
Drosophila
melanogaster
(Meigen)
Musca domestica
(Linnaeus)
Macromastix
zeylandiae
Alexander
Cribellate sheet
web builder
Same as above
Non-sticky sheet
web builder
Non-cribellate,
sticky
orb
web
builder
Same
as
above
Same as above
Non-sticky space
web builder
Same as above
Same as above
Non-sticky sheet
web builder
Non-cribellate, sticky
space web-builder
Same as above
Same as above
Same as above
Same as above
Cribellate orb web
builder
Fruit
fly
House
fly
Crane
fly
None
WO, WOE, WOI, WV,
WVE, WVI
wo,
wo1,
wv,
WVI
wo, WOI, wv,
WVI
None
None
None
wo
None
wo
None
None
None
WO, WOE, WOI, WV,
WVE
wo
WO, WOE, WV
WOI. WVI
WOI, WVI
WOI, WVI
-
s,
L,
VL
s,
L, VL
s,
L, VL
-
-
s,
L
~
s,
L, VL
-
-
~
s,
L
s,
L. VL
S
S
L
VL
sp.
2
w
sp.
I
w
sp.
1
w
Sp.
I
W
&
F
sp.
2
w
sp.
2
w
sp.
2
w
None
sp.
2
w
sp.
2
w
Sp.
2
W
&
F
sp.
2.w
Sp.
2 W
&
F
Sp.
1
W
&
F
None
sp.
2
w
None
None
None
[Notes
to
Table
I
onluring
PURC
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
283
than, the mimetids; and ‘very large’ prey were 1.5-2 times the size of the mimetid (size: estimates of body
volume; see Jackson
&
Hallas, In press
a).
The higher order systematics of spiders is controversial (Levi, 1982), and some of the genus- and
family-placements of species in this study differ considerably in the classifications of Lehtinen (1967), on
the one hand, and Forster,
R.
R.
(1970) and Forster
&
Wilton (1973), on the other. We followed Lehtinen
(1967), but are not in
a
position to provide support for one classification over the other. The controversial
genera are all in Lehtinen’s family Amaurobiidae:
Badurnnu
(Forster,
R.
R.,
1970:
Ixeuticus,
Desidae)
and
Camhridgeu
(Forster
&
Wilton, 1973: Stiphidiidae).
Locomotion on webs and sequences of behaviour during encounters with web spiders will be considered
under the following headings: both densely and sparsely woven cribellate and non-sticky webs, and non-
cribellate sticky webs. Although densely and sparsely woven webs were not discrete classes, but more the
ends of a continuum, behavioural descriptions refer to encounters on very dense or very sparse webs. All
of the cribellate and non-sticky sheet web-builders which were used spun dense webs, if given sufficient
time; sparsely woven webs were obtained simply by removing spiders from webs after a short time. All
non-cribellate sticky webs were sparsely woven; and there were
2
types: orb webs, which are essentially
2-dimensional arrays of regularly-spaced radii and spirals, and space webs, which are 3-dimensional
arrays of threads.
Philoponellu variahilis
is a cribellate orb weaver, and all webs of this species used in
this study were sparsely woven, although this facultatively social spider (Jackson, unpubl. data) will
build other types of webs under different circumstances (for comments on the behaviour of Neotropical
Philoponella,
see Opell, 1979).
Three types of tests were carried out: informal, Type A and Type
B.
The procedures used in each type
of test were essentially the same as those used in tests by the same names carried out in other studies (for
details see: Jackson
&
Hallas, In press
a;
Jarman &Jackson, In press).
Type
B
and
informal
tests
Each type of spider and insect in Table
I
was used in Type
B
tests. In the morning, the mimetid was
put in the cage with
a
spider in its web, observed continuously for
2-4
h, then checked intermittently
during the day, left overnight, and removed the following morning. The prey-types listed in Table
I
were
also used in informal tests which were similar to Type
B
tests but without the fixed test duration and
fixed period of continuous observation. Quantitative results were obtained from Type
B
tests, but
informal tests provided only descriptive and qualitative information.
Since no mimetids were ever observed to capture spiders and insects presented to them in informal
tests in the absence of webs, all tests to be discussed in this paper were conducted by introducing
a
mimetid into a cage containing a web of another spider.
About equal numbers of tests were carried out on each mimetid species, except that only
M.
rnaculosus
was tested with
Inolu
subtilis
Davies. Type
B
tests were carried out using spiders of each of the 3 sizes on
their webs, vacant webs of each type, and webs (either occupied or not by the spider) with eggs present;
tests were carried out with sparsely woven, non-cribellate webs and with densely and sparsely woven
non-sticky and cribellate webs; insects were present in some tests, with or without the spider in the web.
Approximately equal numbers (10-20) of each of these variations on Type
B
tests were carried out, except
that the sample size was increased to
c.
60 each for araneids, theridiids and non-cribellate amaurobiids on
*WO:
web occupied by spider; no eggs
or
insects present. WOE: eggs present in occupied web. WOI: insects present
in cage with web occupied. WV: web vacant (not occupied by host spider); no eggs
or
insects present. WVE: eggs
present in vacant web. WVI: insects present in cage with vacant web
**S,
L,
VL: spider small, large,
or
very large relative to the mimetid
(see
text); size of spiders and insects used in tests
ISp.
I
W:
Mirnefus
sp. observed in nature in web
of
this species
Sp.
I
F:
Mimetus
sp. observed in nature feeding on this species
Sp.
2
W:
M.
rnaculosus
observed in nature in web of this species
Sp.
2
F:
M.
rnaculosus
observed in nature feeding on this species
284
R. R.
JACKSON AND
M.
E.
A.
WHITEHOUSE
sparsely woven webs, with no insects
or
eggs present. This was done in order to obtain more accurate
estimates
of
capture efficiences (see below) against these, the only
3
prey-spiders the mimetids were observed
to capture.
Type
A
tests
Type
A
and
B
tests both lasted
24
h, but there were no periods of direct observation during Type
A
tests, the occurrence of predation being inferred by the presence of dead, dry prey remains at the end of
the test. Since the informal observations indicated that the mimetids caught araneids and theridiids more
readily than other potential prey items, a formal observation procedure was devised with the responses
to
theridiids serving as a standard against which responses to other prey-types were compared.
Also,
the
‘hunger state’ of the mimetid was standardized.
Thirty test series were carried out for each of
5
prey-types (all prey: small), each series consisting of
3
tests. The mimetid was tested with a theridiid on a web during the first and third test. The second test was
with
1
of the
5
prey-types: another theridiid on its non-cribellate, sticky space web; an araneid on its non-
cribellate, sticky orb web; a non-cribellate amaurobiid
on
a
sparsely woven, non-sticky sheet web; a cribellate
amaurobiid on its sparsely woven cribellate sheet web; and
4-8
Drosophilu
rnelunogusster
in
a cage with a
vacant theridiid web.
The test series did not continue if the mimetid did not capture the theridiid
in
the first test. If the mimetid
captured the first theridiid, it was starved for
72
h before the second test began; and when the second test
ended, the third immediately began.
If
neither the second nor third prey item was eaten by the end of the
120
h
(5
day) test series, then the mimetid was regarded as not hungry and the results from this series were
deleted from the statistical analysis. Half of the test series with each type of prey was carried out using
Mimetus
sp.; the other half using
M.
rnuculosus.
Pursuit tendency, capture rate and capture eficiency
The beginning of a pursuit was defined as when the mimetid first vibrated on a web. However, the
mimetids tended to spend long periods completely inactive, which complicated the calculation of pursuit
tendency (percentage of tests during which pursuit occurred) and capture efficiency (percentage of tests
with pursuits during which prey was captured). Periods of inactivity (pauses) between the initiation of a
pursuit and capture of the prey by the mimetid were construed as parts of single pursuits, rather than
breaks between separate pursuits, in order to calculate capture efficiencies (see Discussion).
Often, a mimetid that failed to initiate a pursuit while being observed in a Type
B
test was found to have
captured prey during the interval between intermittent checks. Obviously, mimetids pursued prey during
these tests, but there was no record of how many mimetids had pursued, but failed to capture, prey while
unobserved. Since the presence of the observer may have had an inhibitory effect
on
pursuit initiation
in
some tests, only estimates of capture efficiency were obtained from Type
B
tests. These were calculated
using only the results from tests during which the mimetid was actually observed pursuing the prey before
the capture occurred.
Pursuit tendencies were estimated from the results
of
Type
A
tests by using the capture efficiencies
calculated from the results of Type
B
tests. The mimetids were prone to become unresponsive to prey for
days at a time, apparently as a result
of
satiation. In Type
A
tests, hunger was controlled, and there were
no periods of direct observation; only capture rates (percentage of tests during which prey was captured)
were recorded. Pursuit tendencies were estimated for each type of prey by dividing capture rates obtained
from Type
A
tests by capture efficiencies from Type
B
tests on the same type of prey.
Tests in light-free chamber and tests with painted spiders
During each test, the mimetid was placed in a cage with a theridiid in a web
(8
tests with
Mimetus
sp.;
6
tests with
M.
rnuculosus).
The cage was kept
in
a light-free chamber for
24
h and there %as no direct
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS 285
observation. Four
Mimetus
sp.
and
2
M.
muculosus
were blinded by covering their eyes with opaque enamel
paint, and tests were carried out
48
h later by putting each in a cage with
a
theridiid in a web.
Observations
Observations
in
nature
Most mimetids observed during the day were inactive in the cryptic posture (see below), and
generally they were difficult to locate during the day. The mimetids were easier to locate at night
when they were often active, walking slowly in or near webs.
TABLE
11
Number of Mmetmsp. and Mimetusmacuimmobserved in nature in and not in webs
of
olher spiders
Not in
web
In web
Type
of
web
~~
Mimetus
sp.
18
62
Cribellate amaurobiid
(6),
araneid
(lo),
non-cribellate
Mimetus muculosus
8
100 Cribellate amaurobiid (2). araneid (20), pholcid
(4).
amaurobiid
(I I),
theridiid (35)
pisaurid (24), theridiid (48), uloborid (2)
TABLE
111
Number
of
observations of Mimehs
sp.
and
M.
maculmiw in webs in
nature feeding and vibrating (i.e. plucking, tugging, pulsating, or tap-
ping: see text)
Araneid Theridiid
Feeding Vibrating Feeding Vibrating
Mimetus
sp.
1
5
I
12 4
Mimetus maculosus
8
2
21
4
More mimetids were found in than out of alien webs (Table
11).
Some mimetids were observed
in alien webs performing vibratory behaviours or feeding (Table
111).
One vibrating
M.
maculosus
was observed to attack, capture, and feed on a theridiid (recorded under both ‘feeding’ and
‘vibrating’ in Table
111).
One
M.
maculosus
in a theridiid web was feeding on a theridiid egg sac
(theridiid female not present).
All
other feeding mimetids were eating spiders. The spider species
with which the mimetids interacted are listed in Table
I.
Morphology and appearance
Both species of
Mimetus
were sandy brown spiders, covered with patches of various other
shades of brown.
Mimetus
sp. (adult body length:
4-5
mm) was somewhat smaller than
Mimetus
maculosus
(5-7
mm). The legs
of
both species, especially legs
I,
were long and thin. There were
large conspicuous spines on the inner side
of
the tibia, metatarsus and tarsus of each leg
I
and
I1
of the juveniles and the adult females, but not the males.
286
R.
R.
JACKSON AND
M.
E.
A.
WHITEHOUSE
While at rest, the mimetids adopted a special posture (the 'cryptic posture'): femora
I
and
I1
were angled backwards flush with the carapace; tibiae were angled forward, flush with the femur;
and the metatarsi and tarsi extended along the substratum in front of the spider's body. This
posture was adopted by the mimetid while it sat on branches, twigs, rock ledges, the sides of
buildings and fences, detritus in webs,
or
the silk in densely woven webs. In the cryptic posture,
the mimetid was difficult to detect, especially
if
on a branch or twig. The spider resembled a bud,
and the resemblance was enhanced by two rounded protrusions on the dorsal surface of its
abdomen.
Mimetus maculosus
sometimes spun a rudimentary space web on which it rested when
not in an alien web.
Locomotion
A
walking mimetid, in a web, usually stopped frequently and rarely advanced more than
2-5
cm between pauses; but it commonly advanced more
or
less straight forward, uninter-
rupted, for
20-30
cm
if
not on a web. Whether on
or
away from silk, mimetids generally walked
slowly. They performed rotary probing while walking on webs by slowly waving one
or
two legs
(femoral movement;
0.5-2/s)
so
that the tarsus traced vertically oriented ellipses
or
circles (dia-
meter of long axis:
2-10
mm,
c.
5
mm most common). Mimetids walking on other substrata
often made rotary probes, also;
or
they simply waved legs
I
slowly up and down. If the mimetid
was on a web, usually only
1-3
rotations occurred before the tarsus contacted a thread and came
to rest. Longer bouts were common if the mimetid was not in a web. The mimetids made rotary
probes with legs
I
and 11, but legs
I
probed more often than legs
11,
and a leg
I1
only rarely
probed without the ipsilateral leg
I
probing simultaneously.
The normal walking gait of the mimetid on a web had similarities to rotary probing, the leg
making looping motions to catch threads from the side. Away from webs, the mimetid's loco-
motion sometimes appeared laboured, especially if
it
was crossing a flat surface (e.g. side of a
wooden fence) on which its legs obtained only poor traction. However, the mimetid had little
difficulty walking on thin twigs, beneath which it hung and 'ambulated' along in much the same
way as normally occurred on a thread; and the mimetid tended to move on twigs in preference
to less thread-like substrata, whenever it had a choice.
Elements
of
predatory behaviour
Spreud legs posture
When the spider adopted the spread legs posture (Fig.
l),
all legs were held farther out from
the body than usual
.
Legs
I
were held in the most distinctive position, extending almost straight
out to the side of the body (only
10-20"
forward from perpendicular to the body's sagittal plane).
The femur-patella and tibia-metatarsus joints were flexed only
20"
or less, usually much less.
Usually, the mimetid was standing when its legs were spread, but occasionally, its legs were
spread as it made rotary probes and slowly advanced toward the prey. The mimetid sometimes
performed up-and-down and rotary pulsating (see below) while standing with legs spread. The
adoption of the spread legs posture was most common during, or shortly after, a period when
the prey-spider was active.
Cocked legs posture
In the cocked legs posture, legs
I
and
I1
were held angling forward, roughly parallel
to
each
other and highly flexed (Fig.
2).
Each femur
I
and
I1
angled back alongside the cephalothorax,
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
287
FIG.
1.
Mimetid
performs
rotary pulsating with leg on right in drawing. Note silk strand being pulled by the moving
leg. Legs
I
are in the spread posture. Arrow indicates approximate amplitude
of
movement.
slightly up
(c.
20")
and to the side
(c. 20").
Femur-patella and tibia-metatarsus joints were flexed
c.
90",
so
that the patella and tibia angled up somewhat and the metatarsus and tarsus generally
angled down sharply.
Positioning of legs
I11
and
IV
was more variable, but legs
IV
tended to be positioned
c.
45"
to
the rear. Legs
111
were often held with femora angling out
c.
90"
to the body, and sometimes
slightly down, and with femur-patella and tibia-metatarsus joints flexed
so
as to bring the tarsi
under the body.
Up-and-down plucking
To
pluck, the mimetid moved one
or
more legs rapidly up and down (primarily femoral
movement;
0.5-1
mm,
or
occasionally as much as
3
mm; duration:
c.
0.25
s).
The movement up
was usually
2-4
times faster than the downward return movement. The tip of the tarsus either
remained on the thread during the entire cycle,
or
the thread snapped free as the leg moved up.
Any
of
the spider's eight legs might pluck at any given time, and sometimes all eight legs plucked
simultaneously. Usually, however, plucking was performed with
1-4
of the spider's legs
I
and
11.
Use
of
legs
I1
was the most common. Sometimes, a bout of
c.
four successive plucks occurred,
each separated by a pause of at least
2
s.
288
R. R.
JACKSON AND
M.
E. A.
WHITEHOUSE
FIG.
2.
Mimetid (facing down) with
legs
cocked.
Pedal plucking
While performing pedal plucking, the spider stood (never walked) in the cocked legs posture.
Femora
111
were held about perpendicular to the sides of the body and angling down and to the
side
c.
45".
The femur-patella and the tibia-metatarsus joints were flexed
90-135"
and
c.
20",
respectively.
As
a result, tarsi were arched sharply under the body, sometimes crossing the
midline.
To
pedal pluck (Fig.
3),
the two legs
111
moved simultaneously in alternating phase at
1
-2/s
(bouts:
2-5
s
most common, but sometimes as long as
c.
20
s;
movement smooth), making
vertically oriented circles (diameter:
2-3
mm)
or
ellipses (vertical axis somewhat greater than
horizontal:
2-3
mm
vs.
1-2
mm). Tarsi moved up and forward then down and back to trace the
circle
or
ellipse. Tarsal movement resulted from the femur tracing similar circles
or
ellipses and
simultaneously rotating about its own axis slightly (rotating forward as the leg moved forward
and backward as the leg moved backward), thereby amplifying the tarsal movement. Joints distal
to the femur remained more
or
less stationary.
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
289
As the plucking legs moved, they repeatedly caught hold of silk lines and pulled them for part
of the cycle. The lines always snapped loose before the end of each cycle. Snapping loose occurred
especially as the tarsi crossed over, apparently as a result of the tarsus
or
metatarsus of one leg
moving against and freeing the line being pulled by the other leg.
As
the two legs interfered in
this way with each other's progress, sometimes the movement of one leg was slowed relative to
that
of
the other and the phase relations of the two legs were disrupted. Alternating phase
relations, with one tarsus moving
180"
ahead of the other, tended to
be
restored within the next
one
or
two cycles. The mimetid always remained stationary while plucking (up-and-down
or
rotary).
FIG.
3.
Mimetus
sp. (facing away in the drawing, ventral side up and posterior abdomen toward viewer) performs
pedal plucking with legs
111.
Note
a
silk strand is being pulled
by
each moving leg
111.
Arrows indicate approximate
amplitude of movements.
Tugging
To
tug, the spider moved both legs
I
backward and forward in matching phase (primarily
tibia1 and metatarsal movement;
c.
2
mm;
c.
I/s;
bouts: usually two cycles). Each tarsus
I
remained
on a thread during the entire cycle, causing the silk to move. In each cycle, the first movement was
forward, and there was a momentary pause before the backward motion occurred. Movement
backwards was about twice as fast as the forward movement.
290
R.
R.
JACKSON
AND
M.
E.
A.
WHITEHOUSE
Pulsating
The mimetid pulsated by moving a leg smoothly (i.e. in an uninterrupted, primarily femoral,
motion at more
or
less uniform speed) up and down, with the tarsal tip usually remaining
continually in contact with a thread. Occasionally, the tarsus pushed against the thread during
the downward motion but did not pull the thread up. Two types of pulsating occurred.
The spider performed
up-and-down pulsating
with one leg
I
at a time by moving it up and
down
(c.
2
mm; 1-2/s). Generally,
3-10
cycles occurred in a bout, but occasionally bouts of
50-
100
occurred. Often, the mimetid performed a bout of pulsating with one leg
I,
paused for
c.
0.5
s,
then performed a bout of pulsating with the other leg, and alternated in this fashion
4-5
times.
Occasionally, the mimetid pulsated while walking slowly in a web, but usually it remained
stationary. If the mimetid was stationary, it usually held its legs in the cocked-
or
spread-posture
while up-and-down pulsating. However, while walking, the mimetid often tapped with its
legs
neither cocked nor extended, keeping them instead more
or
less in the normal walking posture.
To
perform
rotary pulsating
(Fig.
I),
the mimetid moved legs
I
as in rotary probing but at
lower amplitude, tracing circles with diameters of only
2-3
mm. Two
or
(most often) only one
leg pulsated at a time. Occasionally, the mimetid walked slowly in the web, alternating between
rotary probing and rotary pulsating (e.g. a probing leg might catch hold of a strand of silk
pulsate, then release the silk and probe again, the mimetid continuing to step the entire time
or
pausing while it pulsated). However, the spider usually remained stationary, in the cocked-
or
spread-legs posture, while rotary pulsating. While standing, the spider sometimes alternated
between bouts of rotary and up-and-down pulsating.
Tapping
To tap, the spider moved its leg and only hit a thread rather than catching and holding it
with the claws as in plucking, tugging and pulsating. To tap, the mimetid moved one leg
I1
rapidly, at low amplitude from side to side
(c.
0.5
mm;
6/s
bouts:
c.
0.5
s; femoral movement
primary).
Lunging
Lunges were initiated from the spread (occasionally)
or
cocked (usually) legs posture.
To
lunge, legs
IV
were extended suddenly, rapidly propelling the spider forward 0.5-1 body length
(duration: less than
0.1
s).
At the same time, the mimetid moved legs
I
and
I1
up and forward,
bringing them over and around the prey to enclose it (Fig.
4).
Next, by flexing its legs, the
mimetid drew the prey towards the chelicerae and bit it.
As
the mimetid’s legs contacted the
prey, sometimes they were seen to begin to flex and draw the prey back toward the mimetid’s
chelicerae while the mimetid’s body was still moving forward. Whether this happened was unclear
in other instances.
Sometimes, legs
111
also were moved forward during the lunge and brought under the prey,
adding a lower layer to the basket formed around the prey by legs I and
I1
(Fig.
4).
Legs
111
did
not always reach the prey, however, when they moved forward during a lunge; and often, instead
of moving forward, they moved backwards along with legs
IV,
contributing to the forward
propulsion of the body.
During the lunge, sometimes the mimetid’s chelicerae were seen to be spread apart with the
fangs extended. Upon contacting the prey with its chelicerae, the mimetid loosened the legs’
grip either immediately (especially if the prey was small)
or
after a few seconds (especially if the
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
29
1
FIG.
4.
Mimetid (facing to the right; ventral side up) at end of a lunge toward and capture of a spider (facing
to
the
left, ventral side
up).
Prey is held by a ‘basket’ formed by legs
1-111
of the mimetid. Note extended legs IV.
prey was large
or
actively struggling). The legs tended to remain around the prey, however, in
a loose ‘basket’. Large and small prey tended to become inactive (apparently paralysed)
5-15
s
and
1-5
s,
respectively, after being bitten by the mimetid.
Post-attack behaviour
The mimetid generally maintained cheliceral contact with, and a loose ‘basket’ of legs around,
the prey for
c.
30
s
after struggling had subsided. Next, if the prey was small, the mimetid usually
carried it to a feeding site without first wrapping it. Larger prey was wrapped before being
carried.
After wrapping, the mimetid walked away, usually going to the edge of the web and eventually
returning to the prey. After returning, the mimetid carried the prey to the feeding site, usually a
sheltered location (e.g. a space beneath a twig) which it had just located (Plate Ic).
The mimetid’s second contact with the prey was similar to the initial attack: it moved its legs
over and brought its chelicerae to the prey (Plate Ia). Its movements, however, were usually
considerably slower than the typical lunging attacks.
Wrapping
The mimetid wrapped by moving legs IV in alternating phase, looping silk over the prey
(0.5-
4/s;
bouts:
30-60
s).
If
the prey was small, legs I11 and usually I1 moved backwards and
forwards, with tarsi on the prey, apparently causing it to rotate. However, the mimetid usually
292
R. R.
JACKSON AND M.
E.
A. WHITEHOUSE'
PLATE. I. (a)-(c) Post-attack behaviour and carrying of prey by
Mimetus
maculosus.
(a) Mimetid (facing down)
returns to a wrapped up prey and scoops it up in a basket formed by legs 1-111. (b) Next, the mimetid (facing up) carries
the prey by walking toward stem at top of photograph, holding prey on a silk line. The line is not visible in photograph.
(c) Having reached a secluded position on the stem, the mimetid (facing to right) feeds. (d)
Mimetus
maculosus
(facing
up) walks on a theridiid egg sac before starting to feed. Another egg sac
is
below and to the left in photograph.
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
293
did not rotate large prey, but wrapped by walking around it. Wrapped prey was covered by a
flimsy coating of silk.
Carrying prey
The mimetids used two methods of carrying prey. The first and second methods were used
primarily with large and small prey, respectively.
1.
While holding the prey with its chelicerae and
legs
I
and
11,
the mimetid walked backwards, stepping with legs
IV,
up a vertical thread.
2.
After
attaching a thread
to
the prey, the mimetid climbed up a thread head-first, stepping with legs
I
and
I1
while legs
111
and
IV
held on to the prey (Plate Ib).
Multiple prey-captures
In some tests, there were juvenile theridiids on the web, with
or
without an adult female present
as well. While the mimetid was feeding on one spider, sometimes a second spider became active.
If the prey on which the mimetid was feeding was small, sometimes the mimetid moved its legs
into the cocked position and performed vibratory behaviours. If the second spider came to within
a body length of the mimetid, the mimetid lunged, simultaneously releasing the first prey, and
capturing the second. The second prey was pulled back against the first, either immediately
or
a
few seconds later, once it was immobilized. Next, the two prey were wrapped together and fed
on as a unit.
As many as three prey were sometimes captured and fed on together. In one instance, a
Mimetus
sp. left a wrapped-up theridiid, on which it had been feeding, walked toward another
theridiid but failed
to
capture it, then returned to its first prey.
Drop on prey
If
there was sufficient thread-free space between the mimetid and another active spider lower
down in the same web, the mimetid sometimes lowered itself slowly on a dragline, with legs
cocked. As,
or
just before, it contacted the prey-spider with its cocked legs, the mimetid lunged.
The mimetid did not perform vibratory behaviour while dropping toward its prey.
Comparison
of
Mimetus
sp. and
Mimetus maculosus
Only
Mimetus
sp. was observed to perform up-and-down plucking, rotary plucking, tapping
and dropping on prey. Otherwise, all described behaviours were performed by both species. Also,
because sequences of behaviour, experimental results, and estimates of capture efficiencies and
pursuit tendencies were very similar for the two species; data were pooled.
Sequences
of
behaviour while on araneid and theridiid webs
Long periods of inactivity were characteristic of the mimetids; when it occurred, activity was
often sudden and brief. Only general trends will be described below.
The lower vertical threads in the non-cribellate theridiid webs are covered by glue droplets.
Normally, prey bump into and stick to these threads, often breaking them free from the sub-
stratum. As a result of the elastic silk contracting, prey may be swung up into the web. Occasion-
ally, a mimetid contacted a vertical thread and stuck to it, but only momentarily, and it never
broke the thread and got swung up into the web. Sometimes, the mimetid climbed up these
threads to enter the web. More often,
it
entered the web from the side
or
the top.
294
R.
'R.
JACKSON AND
M.
E.
A.
WHITEHOUSE
The mimetid entered non-cribellate orb webs by walking across frame and mooring threads to
the orb. The spiral lines within the orb are covered by glue droplets, to which the mimetid
adhered only momentarily,
if
at all. Occasionally, mimetids entered webs by dropping down
slowly on a dragline.
Regardless of the mode of entry, once on an araneid
or
a theridiid web, the mimetid had no
difficulty moving at its characteristically slow pace. Even if the mimetid was artificially dropped
on to an orb web,
so
that it made broad contact with the non-cribellate glue,
it
readily freed
itself.
Predation on spiders
As it walked about in the cage, the mimetid contacted a web, seemingly by accident, slowly
moved on to it, and became inactive, often in the cocked legs posture. This initial pause was
highly variable
in
duration, ranging from a few minutes to
c.
4
h. Sometimes, the host spider
became active, but usually only briefly, as the mimetid walked on to the web.
Also,
the host
spider might become intermittently active during the pause of the mimetid. However, there was
no obvious relationship between the duration of the mimetid's initial pause and host spider
activity.
There was no particular tendency for the mimetid to be facing toward the prey-spider upon
initially entering the web, but when a prey-spider became active, occasionally a quiescent mimetid
would slowly turn and face it. The most common response of quiescent mimetids to activity of
the prey-spider was, however, simply to adopt the cocked leg posture and remain inactive.
Eventually, the mimetid began to vibrate and walk. Only one vibratory behaviour was per-
formed at a time, except that up-and-down
or
rotary pulsating were occasionally performed
simultaneously with pedal plucking. Except for plucking, especially up-and-down plucking,
which was observed only rarely, each of the vibratory behaviours occurred about equally often
but in no discernible pattern.
An araneid
or
theridiid sometimes approached the vibrating mimetid. Occasionally, merely
walking on the web in its normal gait was sufficient to attract the prey-spider to the mimetid.
If
the approach was slow, the mimetid lunged, either as
or
just before the prey contacted its forelegs;
but sometimes the mimetid simply moved its legs over prey that approached rapidly and bit
without lunging.
If the prey-spider was very large
or
approached especially rapidly, the mimetid often dropped
out of the web as,
or
just before, it was contacted.
No
mimetids were killed in any tests, but the
mimetids rarely attracted very large prey-spiders. If large spiders became active, mimetids
usually became quiescent and eventually left the web.
Sometimes the mimetid slowly approached the prey-spider, with
or
without vibrating as
it
walked, occasionally coming close to, lunging at, and capturing it. The mimetid more often
remained in one location and vibrated intermittently. If the prey-spider remained active and
approached, the mimetid persisted at vibrating; but if the prey-spider was not lured, the mimetid
usually became inactive on the web, spending much of the time with legs cocked. As a general
rule, mimetids walked and vibrated during the intervals when the prey-spider was inactive, and
activity by the prey-spider tended to result in active mimetids becoming quiescent.
Sometimes, a mimetid walked on to a web (vacant
or
occupied) and simply stood, often in the
cocked legs posture, and remained inactive for a variable period (e.g.
2
h
or
longer), regardless
of whether the prey-spider became active
or
not. Occasionally, a quiescent mimetid suddenly
lunged at and captured a prey-spider that happened to be walking by. The only response of the
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
295
mimetid before the lunge was to make a slow turn toward the spider except that, if its legs were
not already cocked, the mimetid adopted the cocked legs posture before it lunged.
If not removed after the one-day test period, mimetids tended to remain for several days at a
time in occupied webs and eventually captured the spider. Generally, they did not remain for
comparable periods in vacant webs.
Predation on insects
A
Mimetus
sp. was observed to attack an insect only once. This was an interesting incident in
which the mimetid reacted to a struggling
Drosophila melanogaster,
caught in an occupied
theridiid web (theridiid inactive), as though the
fly
were a spider. While plucking, the mimetid
very slowly approached, lunged at, and bit the fruit
fly.
It immediately carried the
fly
by backing
3
cm higher in the web then abandoned it. The
fly
was still struggling. This was the only test in
which a mimetid was observed to vibrate while pursuing an insect. There was no evidence in any
tests of a
Mimetus
sp. eating insects.
If
a
M.
maculosus
was walking on a web and contacted a living, but ensnared, insect with its
probing leg
I,
it sometimes simply moved forward, after a pause, and fed briefly
(c.
15
min;
normal feeding time on a comparable size spider:
c.
1
h). The mimetid sometimes reacted in the
same way to recently killed and wrapped-up, but unattended, insects in araneid webs.
OOPh
WY
Mimetus
sp. and
M.
maculosus
were observed to eat theridiid eggs once and five times,
respectively, in the laboratory. In each instance, the theridiid was present in the web. Initial
contact with the eggs was seen only once: an
M.
maculosus
walked slowly on to an egg sac
(Plate Id), seemingly by accident and without performing vibratory behaviours
or
lunging. After
c.
45
min of simply standing on the egg sac and intermittently walking about on and near it,
the mimetid began to chew and salivate on the silk and eventually (after
c.
60
min) made a
large hole and began eating eggs one at a time.
Sequences
of
behaviour while on other web-types
Sparsely woven, non-sticky webs
Mimetids readily entered sparsely woven, non-sticky webs and often remained in them for
long periods. They performed vibratory behaviour and sometimes, during sequences with general
similarities to those on araneid and theridiid webs, captured spiders (non-cribellate amaurobiids
only) while on these webs.
No
insects or eggs were eaten, however.
No
mimetids captured pholcids. Six mimetids (four were vibrating; two were simply walking)
were attacked by large pholcids, but each mimetid managed to escape (one lost a pair of legs;
the others suffered no obvious injuries). Pholcids attack by using their exceedingly long legs to
throw silk over their victims, and they can attack while their bodies are still well out of the
mimetid's reach. The mimetid's vulnerability apparently resulted from failure to recognize that
the pholcid was within attack range.
Failed attempts to capture non-cribellate amaurobiids were observed too. Twice, the mimetid
was oriented
c.
45"
away from a small lured amaurobiid and, as the amaurobiid contacted the
mimetid, the mimetid turned and misaimed a lunge. In another five instances, an amaurobiid
(two small; one large; two very large) was lured. It contacted the mimetid head on; but no lunge
occurred, the mimetid dropping out of the web instead.
296
Densely woven and cribellate webs
If the mimetid encountered a sparsely woven cribellate web it usually made numerous attempts
to enter, repeatedly contacting sticky threads, moving away and trying again farther along on
the web. In contrast, the mimetids often moved away from, without attempting to enter, densely
woven webs (cribellate or non-sticky).
If a mimetid entered a densely woven web, it often moved
off
soon afterwards, but occasionally
it adopted the cryptic posture and remained inactive for many hours, either on detritus in the
web or, less often, the silk itself.
No
mimetids were observed to feed
or
perform vibratory
behaviour on cribellate
or
densely woven webs. The locomotion of the mimetid on densely
woven, non-sticky webs sometimes appeared laboured, similar to its locomotion on flat non-silk
surfaces.
The mimetid’s first contact with webs was always with its legs. If a sticky portion of a cribellate
web was contacted, the mimetid’s leg adhered. The mimetid stopped, then lifted its leg
or
backed
away, pulling the legs free.
A
period of grooming usually followed soon afterwards. The mimetid
generally freed itself promptly if only the tips of the mimetid’s legs contacted the cribellate glue.
However, mimetids that were artificially dropped on to cribellate webs and made broad contact
with their bodies or the sides of their legs tended to remain stuck for many minutes or hours at a
time.
R.
R.
JACKSON AND
M.
E.
A. WHITEHOUSE
Pursuit tendency and capture eficiency
The only spiders captured by the mimetids were araneids, theridiids and non-cribellate
amaurobiids (Type
A
&
B and informal tests). Only a few large araneids
(Mimetus
sp.:
1;
M.
maculosus:
4)
and theridiids
(Mimetus
sp.:
3;
M.
maculosus:
2)
were captured.
No
very large
araneids or theridiids were captured.
All
captured non-cribellate amaurobiids were small.
The mimetids readily captured small araneids and theridiids, comparable capture efficiencies
being estimated for these two prey-types, but capture efficiency was considerably less for non-
cribellate amaurobiids (Table
IV).
Estimates of pursuit tendencies for these three prey-types,
however, were similar (Table
V).
Tests in light-free chamber and with painted spiders
Both mimetid species captured theridiids in the light-free chamber and when blinded. The
behaviour of the blinded and normal mimetids was not distinguishable.
TABLE
IV
Results from Type
B
tests
of
mimetids with different types
of
prey
Non-cribellate
Theridiid Araneid amaurobiid
Observed pursuits
43 49 28
Capture efficiency
(%)
9
I
82
14
Capture efficiencies with araneids and theridiids not dis-
tinguishable (tests
of
independence with Yates’ correction). Cap
ture efficiency with non-cribellate amaurobiids less than with
araneids
(x2
=
30.307,
P
<
0.001)
and theridiids
(x2
=
38.318,
P
<
0.001).
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
TABLE
V
Results
of
Type
A
tests
of
mimetidy with different types
of
prey
297
Non-cribellate Cribellate
Theridiid Araneid amaurobiid amaurobiid Insect
No.
of
test series 23
20
21
26
17
Capture rate
(%)
74
70 19
0 0
Pursuit tendency
(%)
81
85
100
-
-
Observations on other mimetid species
We have observed mimetids in New Zealand and Sri Lanka, in addition to
Mimetus
sp. and
M.
maculosus,
occupying alien webs and, in a few cases, feeding on spiders (Table VI). None of
these species has been observed feeding on any other type
of
prey. Observations have been listed
by the name of a town close to, or at which, the observations were made. (For additional
information about each locality: see Jackson
&
Hallas, In press
a).
Voucher specimens, however,
have been deposited with the Otago Museum (New Zealand mimetids) and British Museum (Sri
Lankan).
TABLE
VI
Observations in nature
of
mirneiids oiher than
IMimetcs
sp.
and
Mimetus
macdosus
occupying webs
Location Activity Type
of
web
Galle’ Not feeding Araneid
(2),
theridiid
(1)
Matamatat* Feeding Araneid
(I)
Polonnaruwa* Feeding Araneid
(I)
Not feeding Non-cribellate amaurobiid (9), theridiid
(1)
*Sri
Lanka; **New Zealand
One of the mimetids at Matamata was vibrating in the non-cribellate amaurobiid web it
occupied, and the small amaurobiid began to approach, but prey-capture did not occur.
Another mimetid at Matamata, however, walked along a twig, contacted the edge of an
araneid web, walked without pausing into, then across, the orb to the hub and attacked the
quiescent araneid.
All spiders on which mimetids were observed to feed were small. At Matamata, the theridiid
was
Episinus
sp. indet., which builds a ‘reduced web’ containing only a few threads, and all
amaurobiids were
Cambridgea
sp. indet. At Galle, the theridiid was
Achaeranea
sp. indet. and the
araneids were
Cyrtophora citricola
Forskal and
Cyrtophora
sp. indet. At Polonnaruwa, the araneid
was
Cyrtophora
sp.
indet.
Cyrtophora
webs are non-sticky horizontal orbs with a very fine mesh
(see Lubin, 1973).
298
R.
R.
JACKSON AND
M.
E.
A.
WHITEHOUSE
Discussion
Web invasion, araneophagy and aggressive mimicry
Although the mimetids have been the most celebrated example among spiders of aggressive
mimicry, this study is one of the few detailed reports on the predatory biology of any
of
the species
from this family. Web-invasion is apparently a routine behaviour of
Mimetus
sp. and
M.
rnaculosus:
most individuals observed in nature were in alien webs, and web-invasion took place
readily in the laboratory. On araneid, theridiid and non-cribellate amaurobiid webs, both species
performed special behaviours to which the host spiders sometimes responded as they normally do
to small insects on their own webs. With its vibratory behaviour, the mimetid apparently mimics
its own prey’s prey. Whether either of these mimetid species ever aggressively mimics its prey’s
courtship, as Czajka
(1963)
reported for
Erofurcata,
cannot be ascertained from the results of the
present study, however.
Responding to vibrations, web spiders rapidly and efficiently locate and capture prey that blun-
der into their webs. After initially contacting the web, the mimetid almost always paused before
beginning to vibrate. By remaining at the web edge during the pause and subsequent vibratory
behaviour, the mimetid could gain information about the web and its occupants (e.g. whether
or
not the web was occupied and had cribellate glue, and the location and size of the host spider).
While occupying this position, it could make a hasty escape should this be necessary. By performing
vibratory behaviour that elicited responses from the host spider, the mimetid was able to investigate
whether the web was occupied. Other times, it simply waited for the host spider to provide infor-
mation spontaneously.
The mimetids indirectly manipulated the behaviour of their prey by transmitting deceptive sig-
nals across the web.
By
varying the sequences in which different behaviours were used and, to a
limited extent, the characteristics (e.g. amplitude and speed of appendage movement) of specific
behaviours, the mimetids could transmit a complex array of signals to the host spider, and this
was probably important in increasing the chance of eliciting a response from the host spider.
The repertoire of vibratory behaviours used by
Mimetus
sp. is apparently larger than that of
M.
maculosus,
but the explanation for this difference is not obvious.
Pulsating by mimetids may simulate faint disturbances of the web which are likely to elicit
mild (non-attack) responses from the host spider (e.g. orientation
or
slight movement forward),
and these movements of the prey seemed to enable the mimetid to locate its prey while minimizing
the chance of suddenly being attacked itself.
The mimetids tended to spend much of their time inactive in the webs they invaded. Also, the
mimetids often made sudden attacks on spiders which walked past them. From activity of the
host spider early in the encounter, the mimetid evidently received information about the presence
of potential prey. Failing to capture the prey early in the encounter, the mimetid sat and waited,
having received information indicating that waiting might be profitable. Periods of inactivity
were apparently an integral part of the predatory behaviour of the mimetids. These periods
might be interpreted as part of an ambusher’s ‘pursuit’ of its prey, but only equivocally, since
inactivity was not necessarily a preliminary to prey-capture. However, mimetids generally ad-
opted the cocked-
or
spread-legs posture, instead of the cryptic rest posture, while occupying
sparsely woven, non-cribellate webs. Since cocked- and spread-legs were typical of predatory
sequences, it is apparently justified to view periods of inactivity on these webs as being parts of,
rather than gaps between, individual pursuits by mimetids on webs.
After
a
successful attack, the mimetid wrapped its prey, securing it to the web. Bitten prey
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS
299
that struggled while still weakened from the predator’s venom was unable to break free and
escape. Restraining prey with silk seemed to be important in allowing the mimetid to search for
a sheltered feeding place, and sometimes the mimetid was able to capture a second prey with the
first secured to the web. The wrapping behaviour of the mimetid was very similar to that of
many web spiders (e.g. Robinson
&
Robinson,
1973),
except
it
was much slower and resulted in
a flimsier silk coating.
Spiders are potentially dangerous prey
for
another spider, but the mimetids may tend to be at
an advantage in encounters with other spiders because of their highly potent and apparently
spider-specific venom. Also, a mimetid in an alien web can indirectly manipulate the behaviour
of the host spider
by
acting as an aggressive (vibratory) mimic, and the mimetids fed primarily
on small spiders.
Kleptoparasitism and oophagy
Although the mimetids sometimes fed on insects ensnared in the webs of other spiders (‘klepto-
parasitism’ in the broad sense of the word), araneophagy seems to be
of
greater importance for
these spiders. Similarly, oophagy seems to be an opportunistic feeding tactic of secondary
significance
for
the mimetids. This is an interesting tactic, however, because special behaviours,
which are not widespread among spiders, are needed to open up the tough, tightly spun egg sacs
of
theridiids.
Locomotion
The mimetids walked with ease on sparsely woven, sticky and non-sticky, non-cribellate webs,
but they adhered to cribellate glue and sometimes had difficulty walking on densely woven, non-
sticky webs. The mimetids, along with
Portia
and many space and orb web spiders, practised
rotary probing, moved their legs in a looping gait while walking, and had long, slender legs. These
shared characteristics may be especially suitable for locomotion across widely-spaced threads.
However, the mimetids were apparently unable to adopt a different walking gait which would be
more appropriate on densely woven sheet webs
or
away from webs (e.g. pulling the legs in more
under the body and moving legs more backwards and forwards and up and down rather than
looping them out and in).
Sensory control
of
predatory behaviour
The experimental evidence indicated that vision is of little
or
no importance in the predatory
behaviour of
Mimetus
sp. and
M.
maculosus.
Although prey activity had no apparent effect on
the duration of the mimetid’s initial pause, the mimetid was more likely to remain and initiate
vibratory behaviour or a predatory attack if the web occupant became active. This suggests that
silk-borne vibrations were important in providing information about the presence of potential
prey in webs. Also, silk-borne vibrations are apparently important as cues in prey-location (e.g.
the mimetid would turn towards
a
moving spider), attack initiation (e.g. lunged at active spiders)
and size determination (e.g. ceased to vibrate
if
a large spider became active).
Capture eficiency and pursuit tendency
Although the mimetids readily pursued small araneids, theridiids and non-cribellate amauro-
biids on sparsely woven, non-cribellate webs, they were considerably more efficient at capturing
300
R.
R.
JACKSON AND
M.
E.
A.
WHITEHOUSE
araneids and theridiids than non-cribellate amaurobiids. The reasons for lesser capture efficiency
against non-cribellate amaurobiids are not certain, partly because
of
the difficulty
of
observing
complete predatory sequences, but the findings from this study suggest that the mimetids have
difficulty judging prey location and size on amaurobiid webs. There was no evidence that the
mimetids had locomotory difficulty on the sparsely spun sheet webs, but their sensory systems
are apparently more efficient on some web-types than others; and this is, presumably, related to
differences in the vibratory properties of webs of different designs. Directional errors (orientation
away from lured prey), errors in distance judgement (inaccurate timing of lunge), and error in
size judgement (luring of large and very large spiders) were observed in non-cribellate amaurobiid
but not araneid and theridiid webs.
The relationship between predatory hehaviour and acute vision
Portia
is
behaviourally the most aberrant of the araneophagic web-invading spartaeine salticids
(see Introduction) because it also spins large, prey-capture webs. The biology of
Portia
suggested
a hypothesis (Jackson
&
Blest,
1982)
in which it was proposed that the Salticidae had web-
building ancestors with poorly developed vision, and that acute vision evolved in conjunction
with the adoption of araneophagic invasion of diverse types of webs. Typical web-building
spiders, with their poorly developed vision, rely primarily on web-borne vibrational cues to detect
and locate their prey. Because of structural diversity, spider webs may vary significantly in their
vibrational characteristics, and an araneophagic spider that frequented diverse web-types might
have been faced with a complex array of vibrational stimuli. With the development
of
acute
vision, an araneophagic spider may have been freed from the restrictions imposed by reliance on
vibrational cues and may have been able to become effective as a predator that invaded diverse
types of webs.
Non-visual, web-invading spiders would be expected,
if
this hypothesis is true, to vary consider-
ably in their predatory efficiency on differing web-types; and yet the mimetids (non-visual web-
invaders) were not restricted to just a single web-type. They readily captured araneids (orb
webs) and theridiids (space webs); and their capture efficiences with these two prey-types were
indistinguishable, despite orb and space webs presumably having substantially different vibra-
tional characteristics.
None the less, there is a pronounced difference between the mimetids and
Portia. Porliu
readily
captures, with indistinguishable efficiency (Jackson
&
Hallas, In press
a),
not only araneids and
theridiids on orb and space webs, but also a wide range of other spiders on other web-types. In
particular,
Portia
is highly efficient as a predator when it invades cribellate and densely woven
webs on which mimetids were never observed to catch prey.
On initial examination, these findings are consistent with the hypothesis. If the mimetids had
been found to be highly efficient as araneophagic web-invaders
on
the same wide range of webs
as
Portia,
then the hypothesis would have been placed in jeopardy. However, interpreting the
variance which occurred in predatory efficiencies of mimetids on different web-types is compli-
cated by the necessity of considering both sensory and locomotory factors.
The mimetids stuck to cribellate but not non-cribellate glue and they tended to have difficulty
moving across densely woven webs. Having less secure footing on cribellate and densely woven
webs, the mimetids would probably have more difficulty approaching and formulating an
attack on,
or
escaping when attacked by, the host spider on these webs. And yet they do readily
enter these webs. Perhaps the mimetids often use varied web-types, including types on which
NEW
ZEALAND AND QUEENSLAND PIRATE SPIDERS
30
I
they are ineffective as predators, simply as resting sites. As highly cryptic spiders, they are not
easily seen in cluttered cribellate webs, for example. Also, nocturnal, non-visual predators
might tend to be less likely to encounter a mimetid sitting quietly in a web rather than on
a
twig
or
a stone.
However, the mimetids readily pursued non-cribellate amaurobiids on sparsely woven sheet
webs, and their relative difficulty in capturing this prey-type would seem to be related more to
sensory than to locomotory factors. A more detailed study of predatory sequences on non-
cribellate amaurobiid webs would be valuable. but difficult.
The evolution
of
the Mimetidae
This study has shown that a New Zealand and an Australian mimetid species are araneophagic
spiders which invade alien webs and practise aggressive mimicry. These findings reinforce the
widespread impression that mimetids are a family of web-invading araneophagic spiders. Reports
in the literature (Gerhardt, 1924; Berland, 1932; Bristowe, 1941, 1958; Archer, 1947; Czajka,
1963; Lawler, 1972; Thompson, 1972; Lawrence, 1981; Wise, 1982) and
our
own observations
on
Mimetus
sp.,
M. maculosus
and other species suggest that araneids and theridiids, and perhaps
linyhiids, are the primary prey for most mimetid species. However, careful study of additional
species from this large, cosmopolitan family of spiders is needed. Web-building has been lost
repeatedly in different lineages of spiders, and all araneomorph spiders may have had web-
building ancestors (Jackson, In press). The ease with which the mimetids move across non-
cribellate sticky space and orb webs, despite sticking to cribellate glue and having locomotory
difficulty on densely woven webs, suggests that the ancestors of the mimetids might have been
spiders similar to the theridiids
or
araneids. Heimer
&
Nentwig (1982), in fact, suggested that
the mimetids are closely related to the araneids.
Forster
&
Platnick (l984), however, recently placed the mimetids along with the Archaeidae
and another eight families in the superfamily Palpimanoidea, implying that they are phylogenetic-
ally remote from the araneids and theridiids. The behaviour of most palpimanoid spiders
is
very
poorly known. Some of the archaeids have been reported to be araneophagic (Millot, 1948;
Legendre, 1961; Millot
&
Legendre, 1964), although it is not clear whether
or
not they invade
webs; and there is apparently no knowledge concerning whether they act as vibratory aggressive
mimics. Forster
&
Platnick (1984) suggested that araneophagy may be widespread among palpi-
manoid families. Also, some of the palpimanoids have been reported to build webs, although
details of web-structure and web-use are not available. A comparative study of palpimanoid
spiders, although potentially rewarding, would be difficult because most species are very small
(0.5-2 mm in body length as adults), cryptic residents of rain forest litter and moss.
Financial support was provided by National Geographic Society Grant
2330-8
1,
a Grant-in-Aid from
the Erskine Foundation
of
the University of Canterbury, and grants from the University Grants Committee
of
New Zealand and the Academic Staffing Committee
of
the University of Canterbury. We thank Terry
Williams for photographic processing, Karen Knight for preparing the drawings of spiders, Val Davies,
Fred Wanless and Ray Forster for taxonomic assistance and David Court for assistance in the field in
Matamata. For their valuable comments on the manuscript, we gratefully acknowledge Brenda Poulsen
and Simon Pollard. The New Zealand Ministry of Agriculture provided an import permit.
302
R. R. JACKSON AND M.
E.
A. WHITEHOUSE
REFERENCES
Archer, A. F.
(1947).
The Theridiidae or comb-footed spiders of Alabama.
Pup. Alu6umu Mus.
nut.
Hist.
22
5
67.
Bcrland, L.
(1932).
Les Arachnidcs.
Encycl.
ent.
16.
Bristowe, W.
S.
(1941).
The comity
of
spiders
2.
London: Ray Society.
Bristowe, W.
S.
(1958).
The world ofspiders.
London: Collins.
Curio, E.
(1976).
The ethology ofpredation.
Berlin: Springer-Verlag.
Czajka,
M.
(1963).
Unknown facts
of
the biology of the spider
Ero
furcutu
(Villers) (Mimetidae, Araneae).
Polskie Pismo
Drees,
0.
(1952).
Untersuchungen iiber die angeborenen Verhaltensweisen bei Springspinnen (Salticidae).
Z.
Tierpsychol.
Foelix, R.
F.
(1982).
Biology
of
spiders.
Cambridge, Massachusetts: Harvard University Press.
Forster, L. M.
(1982).
Vision and prey-catching strategies
in
jumping spiders.
Am.
Sci.
70
165-
175.
Forster, R. R.
(1970).
The spiders of New Zealand. Part
111.
Desidae, Dictynidae, Hahniidae, Amaurobiidac, Nicodomi-
Forster, R. R.
&
Platnick, N.
(1984).
A review of the archaeid spiders and their relatives, with notes on the limits
of
the
Forster, R. R.
&
Wilton, C. L.
(1973).
The spiders of New Zealand. Part IV. Agelenidae, Stiphidiidae, Amphinectidae,
Gerhardt, U.
(1924).
Weitere studien iiber die biologie der Spinnen.
Arch. Naturgesch.
90:
85- 192.
Heimer,
S.
(In
press). Notes on the family Mimetidae, with description of a new genus from Australia.
Ent. Ahh.
Mus.
Tierk. Dresden.
Heimer,
S.
&
Ncntwig, W.
(1982).
Thoughts on the phylogeny
of
the Araneoidea Latreille,
1806
(Arachnida, Araneae).
Z.
zool.
Syst.
Evol.
20
284-295.
Jackson, R. R. (In press). Web-building, predatory versatility, and the evolution
of
the Salticidae. In
The evolution
of
spider webs.
Shear, W. A. (Ed.). Stanford: Stanford University Press.
Jackson,
R. R.
&
Blest, A. D.
(1982).
The biology of
PortiuJimhriuta.
a web-building jumping spider (Araneae, Salticidae)
from Queensland: utilization of webs and predatory versatility.
J.
Zool.,
Lond.
196
255-293.
Jackson, R. R.
&
Hallas,
S.
E. A. (In press
a).
Comparative biology
of
Portia ufricunu, P. ulbimunu,
P.
jimbriutu,
P.
luhiutu,
and
P. schultzi,
araneophagic, web-building jumping spiders (Araneae. Salticidae): utilisation of webs,
predatory versatility, and intraspecific interactions.
N.Z.
J.
Zool.
Jackson, R. R.
&
Hallas.
S.
E. A.
(In
press
b).
Predatory versatility and intraspecific interactions of spartaeine jumping
spiders (Araneac, Sakicidae):
Brettus
udonis,
B. cingulutus, Cyrhu ulgerinu
and
Phueucius
sp. n.
N.Z.
J.
Zool.
Jarman. E.
A.
&
Jackson, R. R.
(In
press). The biology of
Tuieriu erebus
(Araneae, Gnaphosidae), an araneophagic
spider from New Zealand: silk utilisation and predatory versatility.
N.Z.
J.
Zool.
Land, M.
F.
(1969~).
Structure
of
the retinae of the principal eyes of jumping spiders (Salticidae: Dendryphantinae) in
relation to visual optics.
J.
exp.
Biol.
51:
443-470.
Land, M. F.
(19696).
Movements of the retinae of jumping spiders (Salticidae: Dendryphantinae) in response to visual
stimuli.
J.
exp.
Biol.
51:
471 493.
Lawler.
N.
(1972).
Notes on the biology and behavior of
Mimetus eutypus
Chamberlin and Ivie (Araneae: Mimetidae).
Noles Aruchnol. Southw.
3
7-10.
Lawrence, R.
F.
(1981).
The cannibal spiders (Mimetidae).
Naturalist,
Leeds
25
4-9.
Legendre, R.
(1961).
Etudes sur les
Archneu
(AranCides).
11.
La capture des proies et la prise de nourriture.
Bull.
Soe.
Lehtinen, P.
T.
(1967).
Classification of the cribellate spiders and some allied families with notes
on
the evolution
of
the
Levi, H. W.
(1982).
Araneae.
In
Synopsis
and clussification
of
living orgunisms:
71 95.
Parker,
S.
(Ed.). New York:
Lubin,
Y.
D.
(1973).
Web structure and function: the non-adhesive orb-web of
Cyrlophoru moluccensis
(Doleschall)
Millot, J.
(1948).
Faits nouveaux concernant les
Archueu
(Arankides).
Mem.
Inst.
Scient. Muduguscur
IA:
3- 14.
Millot, J.
&
Legendre, R.
(1964).
Les Archaea (AranCides) et leur cocon.
C.r. hebd.
S@une.
Acud.
Sci.,
Paris
258:
4835-
Opell, B. D.
(1979).
Revision
of
the genera and tropical American species of the spider family Uloboridae.
Bull. Mus.
en!.
33
229-231.
9
169-207.
dae.
Orugo
Mus. Bull.
3
I
~
184.
superfamily PdlpimanOidCd (Arachnida, Araneae).
Bull. Am. Mus.
nut.
Hist.
178
1-106.
Amaurobiidae, Neolanidae, Ctenidae, Pscchridae.
Otugo
Mus. Bull.
4
1-309.
zool.
Fr.
86:
316 319.
suborder Arancomorpha.
Annls
Zool.
Fenn.
4
199-468.
McGraw-Hill.
(Araneae: Araneidae).
Forma Functin
6
337-358.
4838.
romp.
Zool.
Hurv.
148.443-549.
NEW ZEALAND AND QUEENSLAND PIRATE SPIDERS 303
Robinson, M.
H.
&
Robinson,
B.
(1973). Ecology and behaviour
of
the giant wood spider
Nephila
maculata
(Fabricius)
in
New Guinea.
Smithson. Contr.
Zoo/.
No. 149:
I
75.
Thompson, M. E. (1972). Notes
on
the biology
of
the genus
Mimetus. Notes Arachnol. Southw.
3 19.
Wickler, W. W. (1968).
Mimicry in plants
and
animals.
London: Weidenfeld
&
Nicholson.
Williams, D.
S.
&
McIntyre. P. (1980). The principal eyes of a jumping spider have a telephoto component.
Nature,
Wise,
D.
H.
(1982). Predation by a commensal spider,
Argyrodes trigonum,
upon its host: an experimental study.
Lond.
288.578-580.
J.
Aruchnol.
10
I1
1
116.
