Biol. Rev. (2000) 75. pp. 649-669    Printed in the United Kingdom     ? Cambridge Philosophical Society 549 
The functional significance of silk decorations of 
orb-web spiders: a critical review of the 
empirical evidence 
M. E. HERBERSTEIN1*, C. L. CRAIG23, J. A. CQDDJNGTON4 
and M. A. ELGAR1 
1
 Department of Apology, University of Melbourne, Victoria 3010, Australia 
2
 Museum of Comparative <Wog;)', Harvard University, Cambridge, MA 02138, USA 
3
 Biotechnology Center, Tufts University, Medford, MA 02155, USA 
4
 Department of Entomology, NHB 105, .National Museum of Natural History, Smithsonian Institution, 
Washington, DC 20560-0150, USA 
(Received 21 July 1999; revised 26 June 2000; accepted 28 June 2000) 
ABSTRACT 
~ A number of taxonomically diverse species of araneoid spiders adorn their orb-webs with conspicuous silk 
structures, called decorations or stabilimenta. The function of these decorations remains controversial and 
- several explanations have been suggested. These include: (1) stabilising and strengthening the web; (2) 
hiding and concealing the spider from predators; (3) preventing web damage by larger animals, such as 
birds; (4) increasing foraging success; or (5) providing a sunshield. Additionally, they may have no specific 
function and are a consequence of stress or silk regulation. This review evaluates the strength of these 
     -- ^~- ?-^. explanations-based on the evidence. The foraging function has received most supporting-evidence, derived 
-from both correlative field studies and experimental'-manipulations. This contrasts with the evidence 
provided for other functional explanations, which have not been tested as-extensively. A phylogenetic 
analysis of the different decoration patterns suggests that the different-types of decorations are as 
-evolutionary labile as the decorations themselves: the analysis shows little homology and numerous 
convergences and independent gains. Therefore, it., is possible that different types of decorations have 
different functions, and this can only be resolved by improved species phylogenies, and a combination of 
experimental and ultimately comparative analyses. 
Keywords: Araneidae; Uloboridae; Tetragnathidae; Stabilimenta; Araneae; Behaviour; Phylogeny. 
CONTENTS 
I. Introduction  650 
II. Occurrence and types of web decorations  650 
III. Silk production '.  651 
IV. Variability in decoration patterns  651 
(1) Ontogenetic variation  651 
(2) Population variation  655 
(3) Individual variation  656 
V. Phylogenetic analysis of decoration patterns  657 
VI. Functional explanations       659 
(1)  Mechanical function       659 
* Author for correspondence: M. E. Herberstein, Department of Zoology, University of Melbourne, Victoria 3010, 
Australia (e-mail: m.herberstein@zoology.unimelb.edu.au) 
650 
VII. 
VIII 
IX 
X 
M. E. Herberstein and others 
  661 
(2) Anti-predator function   662 
(3) Improving foraging success  664 
(4) Decreasing web damage          665 
(5) Thermoregulation   665 
Non-functional explanations  666 
Conclusion   667 
Acknowledgments   667 
References  
? i ???J.^ 
I. INTRODUCTION 
Orb-web spiders  (Araneoidea, Deinopoidea)  spin 
circular or ellipsoid webs consisting of a sticky spira 
that   overlays   an   array   of   non-adhesive   radial 
threads. Frame threads, attached to the substratum, 
hold the radials and spiral in place. The mostly two- 
dimensional orb is orientated vertically    e.g.  the 
garden spider Araneus diadematus), horizontally (many 
uloborids) or at an inclination (e.g. tettagnathids). 
Some orb-web spiders spin horizontal webs with 
extensive scaffolding structures (e.g. the tent spider 
Cyrtophora citncola), while otters have lost the web- 
building  habit   altogether   (e.g.   the   bolas  spider 
Durostkhus furcutus).   Although^rb,-webs  are -in- 
conspicuous,  a  number of taxononncally  diverse 
species   (Scharff &   Coddington    1997)   spin  con- 
spicuous silk  designs on  the surface of the wet, 
" McCook     (1889)     cabled    these    silk    stratum 
'decorations';   Simon^^5)   subsequently  intro- 
duced the term 'stabilimenta', and numerous other 
names including 'device', 'adornment' or   struc- 
ture' have been used.oyer the decades   Nentwig 
& Heimer, 1987). The best known example of web 
decorations are those spun by spiders in the genus 
Argiope that consist either of white silk zig-zag bands 
attached to adjacent radii at the web hub, or a silk 
disc at the web centre (Fig. 1). 
The function of these structures has been  the 
subject  of continued  debate  since  their first  de- 
scription, and there has been httle consensus, tor 
example, they may act to camouflage the spider by 
obscuring its outline  (Schoener & Spiller,   1992), 
prevent web damage by larger animals, suchi as birds 
(Eisner & Nowicki, 1983), provide a sunshield for 
the    spider    (Humphreys,    1992)     stabilise   and 
strengthen the web (Robinson & Robinson   1970), 
or   increase   foraging  success   (Craig   &   Bernard 
1990). Alternatively, these decorations may result 
from regulating silk production  (Peters   1993)  or 
reflect  a stressful physiological state  (Nentwig  & 
Rose   1988). These explanations are not mutually 
exclusive (although this is often assumed); it is, in 
fact likely that web deeorations serve different roles 
in different species, and even multiple functions for a 
single species. .       . 
Research into the function of web decorations has 
been taxonomically biased. Most studies have fo- 
cused  on  the linear or cruciate web decorations 
constructed by Argiope spp. or Uloborus spp., and to a 
lesser   extent   the   discoid,   spiral,   tuft   or   debris 
decorations spun by other species and/or juvenile 
spiders (see Table 1). Of a total of 43 publications 
that discuss a possible function for web decorations 
26   were   specifically   concerned   with.   *cweb 
decorations constructed by Argiope spp.  (Table 1) 
and only 21 published reports attempt to test one or 
more functional  explanations.  Furthermore,  only 
three of the functional explanations have been-tested - 
directly   by  experimental   manipulations;   the -re- 
mainder have been examined by either correlative 
.tudies^field observations: The_goal of thiSTJ^w 
is to summarise the research that has investigated .the 
function   of web   decorations   and   to   provide   a 
phylogeny of the various decoration patterns. 
II. OCCURRENCE AND TYPES OF WEB 
DECORATIONS 
We found reports of web decorations in 78 species 
from 22 genera of orb-web spiders (Tab c Q whose 
webs are exposed during the day (Eberhard, 1990, 
Scharff & Coddington, 1997). There are no records 
of decorations in spiders whose webs are exposed 
only at night. Twenty species of spiders in the family 
Uloboridae, 52 species in the family Araneidae and 
six species in the family Tctragnathidae spin weL 
decorations (Table 2). Within the well-known germ. 
Argiope, 21 species were reported to decorate then 
webs (Table 2). Although web decorations are one o 
the most studied phenomena in spiders, the data u 
Table  2   undoubtedly  suffer  from   at  least  thre, 
defects.    Due    to    ontogenetic     individual    an, 
population-level variability (see below)   the etho 
logical occurrence and diversity of web decoration 
Silk decorations of orb-web spiders 651 
\ 
? 
in a particular species may be under-sampled; 
although capable of making a given decoration, it 
may not yet have been observed. Second, not all 
species in groups known to decorate their webs have 
been studied. Third-, some of the observations listed 
in Table 2 are based on reports in old, relatively 
'popular' literature that may not be accurate. 
Silk decorations have been grouped according to 
their basic pattern (Table 3, Fig. 1). The juveniles of 
many species of Argiope spin silks into discs (e.g. 
Nentwig & Heimer, 1987) and spiders in the family 
Uloboridae often spin a decorative spiral at the web- 
centre (Neet, 1990). Adult Argiope spp. add zig-zag 
bands to the web, in either a linear pattern arranged 
vertically across the hub, or a cruciate pattern, 
where up to four zig-zag bands form a diagonal cross 
(Table 3, Fig. 1). Argiope savignyi sometimes spins a 
silk disc, sometimes a cruciate pattern and sometimes 
combines both types (Nentwig & Heimer, 1987). 
The common feature of these types of silk decoration 
is tlvat they are located near the hub of the web, and 
may occasionally extend into the adjacent spiral 
region. By contrast, the silk tufts spun by 
Gasteracanlha and Witica species (Table 3, Fig. 1) are 
placed on the frame threads while some species-of- 
Micrathena place- them throughout the spiral region 
of the web (C. L. Craig, personal observations). 
Some species incorporate organic material, such 
as-bits of vegetation or the remains of captured prey - 
near the centre of the web. These structures have 
also been described as web decorations. When Gyclosa 
caroli reaches reproductive maturity, it incorporates 
sequentially laid egg sacs into a string of pellets 
(Hingston,' 1927; Craig, 1989). Similarly, prey 
_ remains in Nephila, Cyrtophora and Nephilengys are 
mostly attached to the barrier web surrounding the 
orb (Edmunds & Edmunds, 1986). Cyrtophora also 
attach numerous egg cocoons above the web (Elgar, 
Pope & Williamson, 1983; Edmunds & Edmundsr 
1986),-and Cyrtophora citricola incorporates debris in 
the silk retreat above the hub (Kullmann, 1958). 
Phonognatha graeffei attaches a leaf retreat to the web 
structure (Thirunavukarasu, Nicolson & Elgar, 
1996). 
III. SILK PRODUCTION 
Orb-web spiders have up to six different types of silk- 
producing glands from which at least nine different 
types of proteins are produced (Kovoor, 1987). 
Specifically, araneoid spiders produce a minimum of 
eight different proteins that produce three types of 
silk fibroin and three types of protein glue (Craig, 
1997). Their silks are used to produce the frame and 
sticky catching threads of the orb web, sperm webs 
and egg sacs (Foelix, 1992). The decorative silks 
spun by Argiope bruennichi and A. lobata are produced 
in the aciniform glands and are also used to wrap 
prey (Peters, 1993). Like all silks, decorative silks are 
secreted via tiny spiggots. Hence, web decorations 
are actually composed of loosely arranged, indi- 
vidual fibres (Peters, 1993). Individual spiggots are 
found in clusters, or spinning fields, or* each of the 
spinnerets. Comparison of the anterior, posterior 
and median spinnerets reveals that A. lobata has 
almost twice as many aciniform spiggots as Araneus 
diademalus, a non-decorating web spinner (Peters, 
1993). By contrast, the web decorations of spiders in 
the Deinopoidea (e.g. Uloborus walckenaerius and U. 
plumipes), the sister taxon to Araneoidea, use silks 
produced in the aciniform gland as well as the 
piriform glands (Peters,-1993). 
The reflective properties of web decorations also 
differ from other silks incorporated into the web, - 
The decorations spun by A argentata, U. glomosus 
(100% reflective at 370 nm varying to approxi- 
mately 70% reflectance at 640 nm; Craig & 
Bernard, 1990) and Octonoba sybotides (Watanabe, 
1999a), as. well as their catching silks, are ultraviolet 
(UV) reflective {Craig, Bernard & Coddington, 
1994), while the viscid silk spun by A. argentata 
exhibits law reflectivity jn the UV (Craig & Bernard, 
IV. VARIABILITY IN DECORATION PATTERNS 
(1) Ontogenetic variation 
Spiders display considerable "within-species onto- 
genetic variation in decorating behaviour. The most 
commonly reported ontogenetic changes are those of 
juveniles that initially decorate their webs with 
discoid decorations, but .spin linear or cruciate 
decorations during later stadia and when they are 
sexually mature. For example, juveniles of A. savignyi 
(Nentwig & Heimer, 1987), A. flavipalpis (Ewer, 
1972; Edmunds, 1986), A. aetherea (Clyne, 1969) and 
A. keyserlingi (M. E. Herberstein, personal obser- 
vations) usually construct discoid decorations but 
they construct cruciate decorations after they attain 
sexual maturity. Similarly, juvenile A. aurantia and 
A. trifasciata (Tolbert, 1975) spin discoid decorations 
but the adults spin linear decorations. A. argentata, 
however, spin both linear and discoid decorations as 
juveniles but adults spin only cruciate designs (C. L. 
Craig, personal observations). 
652 M. E. Herberstein and othe 
^3 
^ 
?, 
I" 
s 
5 
5 
t- 
9 
0 
c 
-o    S3 
CO  CO  ??^- 
oi ai o to 
CT)   CX>   CT)   1^ 
?i   ??   CM   CT> 
oc 
01 
? 
3 
CO 
?  CM 
-o-> 
a. 3 
o 
? 
c/   t/. ? 
X*  TJ    u 
? <u   u   a 
!3 xl -d 
-a to 
ijj co 
TO      C3      o 
? 
3 
3 
3 
d S> 
2  c 
c
  o 
ft 
33 XI 
if)   CO 
r-~  co 
CTl   CD 
5i 
<*3   o><   CT)    U 
_  ^  -  PQ 
p c 
c 
CM  ___ 
O 
-~J  o 
a o 
.*.   CM 01 
a   c a -ts 
u 
t: ?**    ~ C    ?    9    ? C 
-? -?  c3  c3 ^ -? o   * 
be  u _Q 
&   hp nj 
v <?> co co , 
- CT> ai ai -? 
? CTl en a>   ? 
U   ?'   ??   ?     d Q ?' ??' ^^ rt 
o  o  o rt 
a, 
x 
rt 
C 
u 
C   o 
X    > 
u 
? a a G- > > 
u <u i>   a; 
id C c    C 
>? o o   o 
aj   iu   a_>   aj 
2  o 
<U     O-    U     01 
g J2"q= -r 
3 
0 
> 
-O 
o 
V 
??c 
o 
J3 
o 
V 
c 
ID 
v   ?   O 
- Q. '3 .'S 
>;   rt   d 
V u 
O, O 
-    - ^ 
? tJ ? o   cj o 
3 
u 
y d c 
a. O O x -a s 
oj    >^ "oj  "aj    C 
o .a ^  ? =   u ooooo-cjiiOi; 
o        ^   o 
2    -222 ? ^ 
u        o ~v   a "2 "33 
id- - -Q ^ w ^;-ys 
OJ   ,    rr   a;   a; t>   aj OJ   Jp    U    u u    u 
?? U Ui O ^ (-. 
-^   o -o 2 u ? 
o -a x) 
JH <J= y^ 
^, d 
s 
3 
U 
8 .2 
x J2 i/ aj 
as 
u -a 
-a  <u 
>- u 
2  ? 
2 2 
O   aj 
-O y3 
S ^   d  ^ XJ  XI  TD  X3 
^d^^ddc.-dddd 
X!  XI  XI 
d   d   d 5^ K^5 5^ 
a 
^ 
pe D. 
?a   o* 
I   & 
I. 
?S- o 
^ > > b ^ ^ ?> 
-a 
"^  S 'C   a 
H -2 
po S 
s 
s c 
*o 
_? a ta te 
!< ? 
& C 
-; ^ 
? a M, _^- 
_ g g a _v: 
5 ^ e K G ^> g ^. cr t/ t./ 
i^ ^. 5 ? a a 9 q 
a   ? ^
J 
C    *?*    ? 
?^
J
 o^ o* V> cr  r^1 
a   B ^0  sy) -^ -^  i:   a 
^^>^<0'0^TN^^       ^^^^-^^i^co^o 
? 
3 
u 
3 
?& 
?- 
c 
Silk decorations of orb-web spiders 
653 
O  CO  't 00 
i? i^? i~?        r^ 
tJ! o ra       CT>       -^-~ 
""!. 00 
en c c c 
o o o (A n tn 
?a .a S IS IS IS 
o o o 
Cn   Cn   00  ^  ^  Q* 
-   t^   ^   c^ 
O 
c 
5 p 
?2 3 5 \J tA t/i 
O C c 
1
 o o ^      ^    ^_(      ^-/      U      W 
o 
c4 ,-, 
=o en 
a 2 
o -? 
co 
c c 
o S 
Pi cy5 
~ 
08 ?' 
bp 
~ 
> 9 c X 
2 -a 
u rt 
Z oU 
LO   05 
Ol   CTi CT>   ?? 
" c 
c   o 
g   bo 
a. c 
? ^5 
o -c 
o .-~. 
cy ^ ?? 
"35 ft- 
?O *   ? 
r-    o    D O c/3 IX 
-3 
H 
ro 
oo 
?. 
M M 
O 
be 
B 
2 
CM     ? 
CT)   00 
?. en 
?.   <L) 
ffi pq 
c. 
?~ 
-. 
?1 C1 
DC 
,  0) V 
N 
a 
^2 u 
'Sj 
? 0 
OJ Z CO 
b/ a. 
T) =8 a. 
i> ? ? 
u, 2 V h o c Rl M Jj s UU t4 
a 
! 
o 1 
QJ 
<u   cu   cu   a;   i-i cj 
a a c a -3 a 
o   o  o   o 12 o 
? ? s cu u CD 
> > > u <u u 
U    U    U    1)     U    W     oj g a c c c c c 
o o o o o o o 
ZZ22Z22 
c e a 0 0 ^ ^ ,c 
> 
i- 
c 
V 
a) c a Rj > 
- B 1 > 1 CO B 'C t) 
X c .2 u u xi 0 y ft 0 a. o 
"0 2 i u IU ~ V o c_ C_ g 2 u 
?JO 
o 
(J3 
'5 
u 
V to 
u u u p X ? ?* v 
u ? E u u y Q 
0 o V 
c a 0 a 
e  -S 
^" 
2 
"to 
fe   b   S C3     ?     ?u 
C3 
d 
a. 
CO 
3 
-a 
a 
=: 
| 
Li 
5" 
d 
a 
a 
CO 
3 co a 
CO B 
cu 
5 
CJ 
^3 
Ar
gi
op
e 
pi
ct
a,
 
A.
 
A.
 
re
in
wa
rd
ti,
 
A.
 
A.
 
a
rg
en
ta
ta
, 
A.
 
V
ar
io
us
 
Ar
gi
op
e 
v
a
rio
us
 
e 
a 
b 
En 
a CO a 
_2 
Ne
og
ea
 
sp
. 
V
ar
io
us
 
Ar
gi
op
e 
A.
 
a
u
ra
n
tia
 
Ar
gi
op
e fl
or
ida
, A
 
A.
 
a
pp
en
sa
 
0 
> 
o 
> 
0 
> 1 ?i 0 Ed > 0 > ?JO a 
~ o 
?a 
q 5 
?X)   c 
-5  2 
"5 0 
?a ?-"5 - &S 
?a 3 b   o 
- 
- 
cH    X 
S "3 O    o 
Z Pi 
"3 
- 
> 
bo 
u 
-- 
7. 
CO 
CO tu 
7: T
he
rm
c 
W
eb
 
pr
 
654 
M. E. Herberstein and o 
Table 2. Patterns of web decorations in orb-web spiders (Na = not available) 
Taxon Decoration type Habitat Sour 
Uloboridae 
Conifaber parvus 
Lubinella morobensis 
Octonoba octonarius 
O. sybotides 
O. varians 
Philoponella herediae 
P. republicana 
P. tingens 
P. undulata 
Uloborus americanus 
U. bispiralis 
U. centiculatus 
U. conus 
U. crucifasciens 
U. diversus 
U. filifaciens 
U. glomosus 
U. plumipes 
U. scutifaciens 
Zosis geniculatus 
Araneidae 
Arachnura sp. 
Argiope aemula 
A. aetherea 
A. appensa 
A. argentala 
A. aurantia 
A. aurocinta 
A. bruennichi 
A, catenulata 
A. clarki 
A. doboensis 
A. flavipalpis 
A. fiorida 
A. keyserlingi 
A. lobata 
A. minuta 
A. picta 
A. pulchella 
A. radon 
A. reinwardti 
A. savignyi 
A. trifasciata 
Caerostris sp. 
Cyclosa bifida 
Linear 
Spiral, linear 
Discoid, spiral, linear 
Discoid, linear, spiral 
Spiral 
Discoid 
Linear 
Linear 
Linear 
Linear 
Linear 
Linear 
Linear 
Cruciate 
Spiral, linear 
Linear, cruciate 
Discoid, linear 
Spiral, linear 
Linear, irregular mat 
Discoid, linear 
Debris 
Cruciate 
Cruciate 
Cruciate 
Cruciate 
Cruciate 
Cruciate 
Linear 
Tri-radiate 
Linear 
Discoid, cruciate 
Discoid, cruciate 
Cruciate 
Discoid, cruciate 
Linear 
Cruciate 
Cruciate 
Cruciate 
Linear 
Cruciate 
Discoid, cruciate 
Discoid, linear 
Linear 
Debris 
Forest 
Na 
Na 
Na 
Na 
Forest floor 
Monsoon forest 
understory 
Na 
Tropical forest 
understorv 
Na 
Tropical understory 
Cactus hedges 
Tropical forest 
understory 
Against tree 
trunks 
Na 
Palm leaves 
Urban (campus) 
Na 
Against tree trunks 
Na 
Na 
Na 
Urban (campus) 
Edges, and forest 
gaps 
Grass clearing 
Prairie 
Parkland 
Grassland 
'Among trees' 
Between trunks 
of palms 
Tree trunks 
Lower shrub 
Na 
Shrub 
Na 
Na 
Na 
Mango trees 
Over river 
Lubin etal. (1982) 
Lubin (1986) 
Peaslee & Peck (1983) 
Watanabe (1999 a) 
Yaginuma (I960) 
Opell (1987) 
Opell (1979) 
Opell (1979) 
Lubin (1986) 
Cornstock (1912) 
Lubin etal. (1982) 
Hingston (1927) 
Lubin etal. (1982) 
Hingston (1927) 
Eberhard (1973)    * 
Hingston (1927) 
Cushing & Opell (1989) 
Marples (1969); Peters (1993) 
Hingston (1927) 
Lubin (1986) 
N. Scharfl" (pers. comm.) 
Robinson & Robinson (1974) 
Elgar el al. (1996) 
Kerr (1993) 
Craig & Bernard (1990) 
Tso (1998 a) 
Robinson & Robinson (1980) 
Malt (1993) 
Hingston (1927) 
Hingston (1927) 
Na 
Na 
Prairie 
Na 
Na 
Robinson & Lubin (1979) 
Edmunds (1986) 
Eisner & Nowicki (1983) 
Herberstein el al. (2000) 
Robinson & Robinson (1974) 
Yaginuma (1960) 
Robinson & Robinson (1974) 
Marson (1947a) 
Levi (1983); Robinson & Robinsoi 
(1980) 
Robinson & Robinson (1974) 
Nentwig & Heimer (1987) 
Tso (1998 a) 
J. A. Coddington (pers. obs.) 
Bristowe (1941) 
Silk decorations of orb-web spiders 
655 
Table 2. (cont. 
Taxon 
C. caroli 
C. conica 
C. centrifasciens 
C. cylindrifasciens 
C. insulana 
C. micula 
C. turbinata 
Cyrtophora bifurcata 
C. citricola 
C. hirta 
C. moluccensis 
Gasteracantha 
brevispina 
G. cancriformis 
G. curvispina 
G. germinata 
G. minax 
Gea eff 
Isoxya cicatricosa 
Micrathena 
sexspinqsa 
Neoscona 
domiciliorum 
N. arabesca 
Neogea egregia 
Parawixia tuberculata 
Singa sp. 
Witica crassicauda 
W. crassispina 
ZUla sp. ] 
ZHIa sp. 2 
Tetragnathidae 
Nephila clavipes 
M. edulis 
?N. maculata 
N. plumipes 
N. tetragnathoides 
Nephilengys cruentata 
Decoration type 
Discoid, spiral, 
debris 
Linear, debris 
Debris 
Silk cylinder 
Discoid, spiral, 
debris 
Spiral 
debris Linear, 
Debris 
Debris 
Debris 
Debris 
Debris 
Linear, tufts 
Tufts 
Tufts 
Linear, tufts 
Cruciate 
Linear 
Linear 
- 
Linear 
Linear 
Linear 
Debris 
Na 
Linear 
Linear, tufts 
Linear 
Linear 
Linear 
Linear, debris 
Linear (juveniles) 
Linear, debris 
Linear, debris 
Debris 
Habitat Source 
Na 
Shrubs 
On trees 
Mangroves 
Rubber trees 
Na 
Na 
Na 
Na 
Na 
Na 
Na 
Na 
Under bushes 
Na 
Shrub 
Tall grass 
Grass 
Na 
Na 
Na 
Na 
Na 
Na 
Na 
Na 
Bushes, forest edge 
Na 
Forest edges 
Forest 
Various 
Mangroves 
Coconut plantations 
Urban 
Levi (1977) 
Bristowe (1941) 
Hingston (1927) 
Hingston (1927) 
Marson (1947*) 
Bristowe (1941) 
Rovner (1976) 
Levi (1977) 
Kullmann (1958) 
M. A. Elgar (pers. obs.) 
M. A. Elgar (pers. obs.) 
Hingston (1927) 
Marples (1969) 
Edmunds & Edmunds (1986) 
Simon (1895) 
Mascord (1970) 
Levi (1983) 
Edmunds & Edmunds (1986) 
Nentwig & Heimer (1987) 
McCook (1889) 
McCook (1889) 
Levi (1983) 
Yaginuma (1960) 
Wiehle (1929) 
Levi (1986) 
Nentwig & Heimer (1987) 
Robinson & Robinson (1980) 
Robinson & Robinson (1980) 
Robinson & Robinson (19734) 
M. A. Elgar (pers. obs.) 
Robinson & Robinson (1973a) 
M. A. Elgar (pers. obs.) 
M. A. Elgar (pers. obs.) 
Edmunds & Edmunds (1986) 
: 
. 
Ulobonds also show ontogenetic change in web- 
building behaviour. Immature Uloboms conus built 
inear decorations,   while  adults  rarely  built  any 
(Lubinid ai,   1982).  Zosis gemculatus adults con- 
structed disks more than linear decorations, while in 
~?srdecorat,ons were m?re **?? 
(2) Population variation 
The proportions of individuals that decorate their 
webs can vary greatly between different populations 
For imtancc A. Jlavipalpis in West Africa (Edmunds, 
y?b   and U. diversus in Arizona, USA (Eberhard, 
197J)   almost   always   spun   web   decorations    By 
contrast    only  25?/o   of individuals  of A.  appensa 
observed  on   Guam  constructed  web  decorations 
(Hauber,   1998),   while   on   neighbouring   Pacific 
islands the frequency of web decorations in A. appensa 
varied   between   3.6?/0   and   76%   (Kerr,   1993) 
Similarly, the proportion of A. argentata that never 
656 M. E. Herberstein and oth 
Table 3. Summary of the types of web decorations and their possible location on the web (see Fig. 1 for schematic 
representations) 
Type Location on web 
Silk 
Discoid Hub 
Spiral Hub 
Linear Hub 
Cruciate Hub 
Tufts Orb. frame 
Debris 
Prey remains Orb, barrier web 
Egg sacks Orb, frame, barrier 
web 
Example 
Juvenile Argiope argentala 
Lubinella morobensis 
Argiope trifasciata 
A. argentala 
Various Gasleracantha spp. 
Mephila plumipes 
Cyrtophora citricola 
Nentwig & Heimer (1987) 
Lubin (1986) 
Horton (1980) 
Craig & Bernard (1990) 
Edmunds (1986) 
Austin & Anderson (1978) 
Edmunds & Edmunds (19{ 
discoid 
Fig. 1. Schematic representation of various web decoration types in orb-web spiders. 
decorated their webs ranged fronT65 % in Panama 
(Robinson & Robinson, 1970) to between 58% and 
78% on Santa Cruz Island, Galapagos (Lubin, 
1975). These data may either reflect facultative 
changes in the decorating behaviour according to 
local environmental conditions, or indicate a genetic 
basis for web-decorating behaviour (Edmunds, 
1986). 
(3) Individual variation 
Individuals of the same species display different 
decorating behaviours that may be influenced by 
changes in local physical factors. For example, A. 
aetherea spun more web decorations in dim light than 
in bright light (Elgar, Allan &  Evans,  1996). By 
contrast, U. diversus constructed more circular ai 
linear decorations following nights of bright i 
lumination compared with nights of low light levt 
(Eberhard, 1973). Food availability also affec 
decorating behaviour. For example, A. aurantia,. 
trifasciata (Blackledge, 1998 A) and A. keyserlh 
(Herberstein, Craig & Elgar, 2000) construct mo 
and larger web decorations when they are mail 
tained on a high-energy diet than when maintainf 
on a low-energy diet. Octonoba sybotides was moi 
likely to construct spiral decorations when foe 
deprived while food-satiated individuals tended I 
form linear decorations (Watanabe, 19996). Fu 
thermore, when A. keyserlingi was fed the san 
amount of prey, the number and size of decoratioi 
was greater when prey encounter rates were ui 
Silk decorations of orb-web spiders 657 
predictable than when they were constant 
(Herberstein et al., 2000). The effects of feeding 
regimes were less pronounced for A. trifasciata 
foraging in the field, which suggests that other 
environmental factors, in addition to nutrition, 
influence the decorating behaviour (Tso, 1999). 
Finally, experimental data suggest that selection 
may favour unpredictable decorating behaviour. 
For example, the stingless bee, Trigona fluviventris, 
learned to avoid the webs of A. argentata that were 
decorated identically over successive days. When the 
orientation of decorations was varied, however, 7". 
^ fluviventris were more likely to be intercepted and 
hence captured in the web (Craig, i994<z;6). 
Correlative field studies further demonstrated that 
when a population of webs spun by A. argentata were 
decorated identically over three successive days, they 
intercepted fewer prey on days 2 and 3 compared 
with day 1. However, when webs in a second 
population were decorated randomly, they inter- 
cepted the same number of insects on each of the 
three days of the experiment (Craig, 1994A, b). By 
contrast some A. flavipalpis spun the same pattern 
more often than expected, while other individuals 
frequently changed the decoration pattern on a daily 
basis (Edmunds, 1986). 
V. PHYLOGENETIC ANALYSIS OF 
DECORATION PATTERNS 
We analysed, the different patterns of decorations (as 
defined by Fig. 1) on a composite phylogeny (Fig. 2) 
of orb-web spiders produced by melding the latest 
relevant phylogenies from the literature. The phy- 
logeny of Uloboridae. is taken from Coddington 
(1990), of Araneidae from Scharff & Coddington 
(1997), of Tetragnathidae from Hormiga, Eberhard 
& Coddington (1995), of Orbiculariae from 
Griswold et al. (1998), and of Entelegynae from 
Griswold et al. (1999). Two characters, presence, or 
absence of decorations (Fig. 2) and type of dec- 
oration patterns (see Table 2), were then traced at 
the generic level using MacClade 3.0 (Maddison & 
Maddison, 1992). 
The first character is binary, and the second was 
treated as unordered. We made no a priori 
assumptions about any phylogenetic transformation 
series linking different types of decorations. A 
number of genera in Table 2 exhibit more than one 
type of decoration, thus making those terminals in 
Fig. 2 'polymorphic'. We checked MacClade's 
interpretation of polymorphic taxa by adding as 
many dummy terminals as necessary within each 
genus to represent the various decoration types 
monomorphically. The lack of species-level clado- 
grams for genera exhibiting diverse patterns is 
unfortunate. For example, if Nephila tetragnathoides 
were basal to the remaining Nephila in Table 2, the 
presence of a debris decoration in Nephilengys would 
make the debris pattern primitive for Nephila. 
Contrariwise, if N clavipes and N. maculata were 
basal, the linear pattern would be inferred as 
primitive. Similar effects are possible in Argiope, 
Uloborus, Cyclosa, and Gasteracantha. 
For the Uloborids (Fig. 3A) these data imply that 
linear or discoid decorations are primitive because 
only these two patterns are present in each of the 
- four web-decorating genera in the cladogram. Spiral 
or cruciate patterns, such as in U. crucifasciens or U. 
diversus (Table 2) evolved de novo. The debris pattern 
is reconstructed to be primitive for Nephila (Fig. 3B) 
by outgroup comparison to Nephilengys. In Witica 
(Fig. 4A), tufts, debris, or both may be primitive, 
but the debris pattern is de novo to Cyrtophora. In 
Neogea and Gea, the patterns are monomorphic, and 
thus either the linear or cruciate pattern is plesio- 
morphic for Argiope; discoid and 'tri-radiate' (Table 
2) are reconstructed as de novo in Argiope. However, 
the polymorphic Argiope species tend also to be the 
better known ones. Further observations may reveal 
that currently monomorphic genera such as Gea may 
also be polymorphic. Linear decorating patterns are 
de novo in Micrathena and Caerostrisr^but could be 
homologous and plesiomorphic for Gasteracantha, by 
outgroup comparison to Isoxya (Fig. A-A). Finally, all 
decorating patterns are de novo in Cyclosa (Fig. 4B). 
Considering just the presence or absence of web 
decoration at the generic level (Fig. 2), these data 
- require at least nine gains and one loss to explain the 
distribution of web decorations within Orbiculariae 
(Scharff and Coddington, 1997). Compared to other 
comparative data on orbicularians the fit of 'web 
decorations' to the cladogram is worse than any 
other character used by Scharff and Coddington 
(1997). In other words, there seem to be frequent 
evolutionary changes in web decorations. This 
pattern more or less repeats at the level of decoration 
patterns - states are far more likely to be convergent 
between genera than homologous. Although without 
phylogenies it is impossible to be precise, the diversity 
of patterns within genera will also require many 
hypotheses of convergence to be explained. This 
high degree of lability in the evolution of web 
decorations certainly supports the possibility that 
different  types of web decorations have different 
658 
M. E. Herberstein and oth 
*s 
?| 
<J o 
Q   3 
ca Stegodyphus 
a Megadictyna 
=3 Dictyna 
=3 Badumna 
KM Amaurobius 
?3  Tangelta 
ICJ Zoropsis 
Psechrus 
Tapinillus 
Deinopis 
Menneus 
Tangaroa 
Waitkera 
Polenecia 
Siratoba 
Ariston 
Sybota 
Orinomana 
Hyptiotes 
Miagrammopes 
Ponelta 
Purumitra 
Daramulunia 
Uloborus 
Zosis 
Philoponella 
Octonoba 
Phonognatha 
Clitaetra 
i Nepbila 
Herennia 
i Nephilengys 
lO A.-ili.i 
Dolichognatha 
Meta 
MeteMna 
Chrysometa 
Leucauge 
Tetragnatha 
Pachygnatha 
Gtenognatha 
lO Theridiosoma 
Mysmena 
Patu 
Anapsis 
Pimoa 
Q Linyphia 
Microlinyphia 
Synotaxus 
Isicabu 
Tekella 
Nesticus 
Gaucelmus 
Steatoda 
Anelosimus 
Chorizopes 
Witica 
Arachnura 
Cyrxophora 
Mecynogea 
Neogea 
Gea 
Argiope 
Mastophora 
l Cyrtarachne 
i Pasilobus 
i Hypognatha 
Archemorus 
Arkys 
i Encyosaccus 
i Xylethrus 
Chaetacis 
i Micrathena 
1 Aspidolasius 
? Caerosths 
Gastroxya 
> Augusta 
1 Macracantha 
? isoxya 
Austracantha 
Togocantha 
Aetrocantha 
? Gasteracantha 
a Scoloderus 
a Acanthapeira 
Hyposinga 
m Singa 
a Zygiella 
Kaira 
=a Metapaira 
Dolophones 
Anepsion 
Colphapeira 
o Nuctenea 
?i Cyclosa 
? Araneilla 
ea Eriophora 
a Varri/coss 
a IWanff0f3 
Neoscona 
Q Larinia 
Aculepeira 
i Araneus 
i Cercidia 
l Pronous 
i Metazygia 
? Alpaida 
i Enacrosoma 
i Bertrana 
jEustala 
Wixia 
Acacesia 
CO 
J3 bp 
3 
cr 0 
c 
d 
o 
c 
o 
L 
"3 
tf 
s 
a- 
o 
bo | 
-a 
o G 
M 
IS 
o 
U 
a 
? 
Q 
o 
-a 
U 
Silk decorations of orb-web spiders G59 
(A) 
(B) 
Fig. 
Sec 
Decoration patterns 
unordered 
I      I absent 
r?x^1 linear 
spiral 
discoid 
?? cruciate 
?? tufts 
liiitil polymorphic 
1
       | equivocal 
3. Cladogram tracing decoration patterns on equally weighted tree for Uloboridae (A) and Tetragnathidae (B). 
text for further details. 
functions. Moreover, it is likely that the same type of 
decoration differs in function between species. 
behaviour. We discuss each in turn, together with a 
critical analysis of'the available evidence. 
VI. FUNCTIONAL EXPLANATIONS 
The evolutionary significance of decorating behav- 
iour has attracted considerable debate, and there are 
several general explanations for the function of this 
(1) Mechanical function 
Simon (1895) was the first to ascribe a mechanical 
function to the cruciate decorations constructed by 
A. argentata, suggesting that the silk ribbon 
'strengthens the position occupied by the spider". 
-.- 
660 
? 
M. E. Herberstein and 0| 
(A) 
o 
-c 
a o 
t 
O 
? 
o 
e 
o i 
a to 5 s 5 D 
o 
0) 3 ? 
2 tt < 
o 
-c 
o 
I 
<n n: n n n TO TO TO 
c 
?c u 
e 
G 
-Q 
o 
s 
5 
TO 
c 
o 
| 
3 c 0 
c 
-c 
u o TO 
? 
o 
o C 
a 
"2, 
43 
o TO 
0) TO 
-C tj 
TO C 
? 
u 
5 
.3 
s TO 
1 
5. 
w 
C 
to 
8 
(3 
H 
8 
TO o 
TO 
s 
c TO 
1 | o -2 
I 
TO ? 
C TO a 6 
a (2 
TO ? 
e TO 
O I 
s 
TO 
3 D ? n a E3 ? n G ? E3 i 
(B) 
Decoration patterns 
unordered 
absent 
E22D  linear 
^^ spiral 
discoid 
BH cruciate 
?OB debris 
wm tufts 
HB polymorphic 
t=l equivocal 
TO s; O o </> 
?S -C S 
c 
CO I 
,^ c ? QJ o 
-* Uj s 
13 ? ? D 
Decoration patterns 
unordered 
I      I absent 
l:::::::::l linear 
W7\ spiral 
HU1] discoid 
?? cruciate 
BB debris 
?? tufts 
ETS5I polymorphic 
1=1 equivocal 
Fig. 4. Cladogram tracing decoration patterns on an equally weighted tree for Araneidae. The tree is split into 
parts (A, B) for clarity. See text for further details. 
Accordingly, the silk bands may enable the spider to 
adjust the mechanical state of the completed web 
(Robinson & Robinson, 1970). Therefore, the 
variability in web decorations may reflect differences 
in the amount of mechanical adjustment require 
strengthen the web (Robinson & Robinson, 19 
For example, Robinson & Robinson (1# 
attributed a mechanical strengthening functid 
Silk decorations of orb-web spiders 661 
the  zigzag  bands   on   the   moulting   platforms   of 
Nephila clavipes. 
This idea has been tested indirectly by comparing 
the frequency with which decorations are added to 
webs at sites characterised by different wind con- 
ditions. The proposed role of decorations is to 
stabilise webs in windy sites, where they are likely to 
require 'mechanical support'. Hence, more 
decorations should be added to webs in windy sites 
than to webs spun under calmer conditions. These 
observations provide little supporting evidence; 
there was no difference in the frequency of 
decorations.spun by A. argentata in habitats exposed 
to strong winds and those in sheltered habitats - 
(Lubin, 1975;Nentwig&Rogg, 1988). Nevertheless, 
Neet (1990) reports an increase in the frequency of 
circular decorations in Cyclosa insulana on windy 
days. However, these observations were conducted 
on three separate days, two of which were classified 
as 'calm' and data on wind speed (or any other 
environmental variables including rate of prey 
capture) are not provided (Neet, 1990). Wind speed 
manipulated in the laboratory had no effect on the 
decorating behaviour of A. argentata (Nentwig & 
Rogg, 1988). 
A mechanical function of web decorations seems 
unlikely, although the idea has not been tested 
extensively. For example, it is not known whether ~ 
decorated webs are more stable than undecorated 
webs, nor is it clear how web decorations could 
stabilise or strengthen the YVjeb. Indeed, other types 
of web decorations, particularly debris or tuft: 
decorations (Fig; 1), seem an unlikely strengthening 
tool because the decorating silk is.generally laid as an 
unstressed mass and often only attached to a single 
thread (Eberhard, 1973,. 1990). Perhaps the 
strongest supporting evidence will come from com- 
parative studies that compare the mechanical stab- 
ility of webs constructed by decorating and non- 
decorating species. 
(2) Anti-predator function 
Web decorations may protect spiders from predatory 
attacks. There are thought to be several different 
protective mechanisms that depend, in part, on the 
various types of decorations. Pellets of prey remains 
or egg sacks may act as a decoy, confusing the 
predator who attacks the pellet rather than the 
spider (Hingston, 1927). Spiral or discoid 
decorations may conceal and hide the spider from 
predators (Eberhard, 1973; Ewer, 1972) and cru- 
ciate or linear bands may change the apparent shape 
of the spider, making it less obvious to predators 
(Edmunds, 1986). Cruciate decorations may also 
increase the apparent size of the spider by appearing 
as extensions of pairs of legs, and thus protect it from 
gape-limited predators (Schoener & Spiller, 1992). 
Web decorations, including the silk tufts used by 
Gasteracantha spp., may protect spiders indirectly by 
advertising the presenceof the web, which may be a 
negative stimulus to some avian predators (Horton, 
1980). 
The anti-predator function for web decorations 
has been examined in a number of ways. First, 
variation in web decorations has-been interpreted to 
"- reflect the relative abundance of predators. For 
example, the frequency of decorating A. argentata on 
different islands in the Galapagos is associated with 
the number of potential predators (Lubin, 1975). 
Spiders on Daphne Island, which lacks potential 
spider predators, only rarely spin web decorations, 
whereas spiders spun decorated webs more fre- 
quently on Santa Cruz Island, which maintains a 
higher diversity of predators such as birds and 
lizards. However, there are no reports of direct 
predatory attacks by any of the suggested predators 
(Lubin, 1975). It is also possible that there may be 
?other differences between the two islands that affect 
decorating behaviour. 
U A survey of A. argentata in the Bahamas showed 
that cruciate decorations were mostly spun by 
medium-sized spiders" hut  not  by small  or large 
.spiders (Schoener & Spiller, 1992). A. argentata may 
~ use cruciate decorations to increase their apparent 
size to predatory lizards (Anolis spp.), whose gape 
size places an upper limit on the size of their prey. 
Consequently,  medium-sized  spiders:"benefit  most 
-from spinning decorations but small spiders could 
not increase their apparent size beyond the gape size 
of the lizard artd large spiders are already large 
enough (Schoener & Spiller, 1992). While this 
explanation is certainly possible, the mortality rate 
due to predatory attack of differently sized A. 
argentata was not measured nor was the response of 
lizards towards decoration size tested experimentally 
(Schoener & Spiller, 1992). 
Finally, the variation in web decorating may 
reflect a trade-off between foraging efficiency and 
predator avoidance (Blackledge, 1998 6). Thus, well- 
fed spiders might invest more energy into predator 
avoidance by spinning more decorations compared 
with food-deprived animals. Laboratory experi- 
ments revealed that food-deprived A. trifasciata 
constructed fewer web decorations than well-fed 
spiders (Blackledge, 19986). 
662 M. E. Hcrberstein and other* 
A more direct approach attempts to demonstrate 
experimentally that the risk of predation is reduced 
by decorations. Captive blue jays (Cyanocilla cristata) 
were allowed to choose individuals of A. argentata and 
A. trifasciata without a web, spiders in undecorated 
webs and spiders in decorated webs. The birds 
attacked spiders without a web most frequently, and 
spiders in undecorated webs more frequently than 
those in decorated webs (Horton, 1980). These data 
are consistent with the anti-predator function, but 
the decorations may not have acted to camouflage or 
conceal the spider, but rather the birds may have 
learned to associate web decorations with a negative 
stimulus - a sticky web (Horton, 1980). 
The anti-predatory function of decorations is 
widely cited (Table 1) despite the paucity of 
experimental evidence. The absence of decorating 
behaviour in nocturnal species is consistent with the 
anti-predator explanation because few visual 
predators are likely to be active at night. However, 
records of the prey of diurnal sphecid predators 
reveal no consistent patterns. For example, the 
predominant genera captured by Sceliphron laetum in 
Madang, PNG were Argiope, Gasteracantha and 
Neoscona (Elgar & Jebb, 1999); only the latter is not 
known to decorate webs. By contrast, S. madra- 
spatanam conspicillatum rarely captured Argiope spp., 
but preyed more frequently on species ofAraneus and 
Neoscona '(Adato-Barrion & Barrion, 1981). Of 
"course, it is difficult to interpret these data without 
knowledge of the abundance and accessibility of the 
different genera. 
However, the possibility that predators may also 
learn to associate web decorations with prey location 
is a strong argument against an anti-predator 
function (Robinson & Robinson, 1970). In a held 
experiment, crickets attached to screens were offered 
as prey to wild birds in the field. The crickets were 
placed in the centre of a 'model stabilimentum' 
consisting of a white zig-zag stitching arranged in a 
cruciate pattern. Eventually, the birds showed a 
preference for crickets with models over crickets 
without models (Robinson & Robinson, 1970). 
(3) Improving foraging success 
There are essentially two mechanisms by which web 
decorations may increase the foraging success of 
web-building spiders. First, they may attract pol- 
linating insects by reflecting UV light in patterns 
similar to UV markers on flowers. Second, UV 
patches created by web decorations may indicate 
gaps in vegetation, which elicits flight behaviour in 
many insects (Craig & Bernard, 1990). A foraging 
function has been tested extensively using spider 
that construct cruciate, linear and discoi< 
decorations (eg. Craig, 1991; Craig & Bernard 
1990; Hauber, 1998; Tso, 1996, 1998 a, * 
Watanabe, 1999a; sec Table 1), and there are tw< 
lines of supporting evidence. First, the presence ant 
the number and size of decorative bands wen 
related to the presence of prey and to prei 
interception rates of the web. Second, more direc 
studies using choice experiments and field experi 
ments showed that insects approach UV-reflectin; 
web decorations more frequently-than webs withou 
UV-reflecting decorations. 
The orientation and presence of decorative sill 
on webs spun by naturally foraging A. argentata wet 
manipulated, and the prey-capture rates wet 
estimated by recording the web-damage patten 
caused by prey interceptions (Craig & Bernan 
1990). Decorated webs intercepted significant! 
more insects than undecorated webs and the pre 
ence of the spider itself, which is characterised by 
reflecting abdomen, further enhanced pre"V-captui 
rates. Moreover, a within-web comparison showt 
that the vertical web half containing the decorati\ 
band also intercepted more prey than the ui 
decorated web half (Craig & Bernard, 1990). 
A further field study measured the prey-inte 
' ception rates at A. argentata webs under thn 
conditions: interception rates of solitary spider web 
interception rates of webs within 3 m of at least or 
other web, and interception rates of webs with boi 
the spider and the -decorations removed (Crai 
1991). Webs of A. argentata foraging within 3 in 
each-other intercepted more insects than solitar 
foragmg individuals.- Furthermore, within the 
clusters, decorated webs captured more prey tha 
undecorated webs (Craig, 1991). However, wht 
the spiders and the decorations were remove 
solitary webs intercepted insects at similar rates 
webs in a cluster (Craig, 1991). 
The presence of web decorations was also relati 
to prey interception in A. trifasciata (Tso, 199( 
Decorated and undecorated webs did not differ 
size, but decorated webs intercepted more flyii 
insects than undecorated webs. However, the w 
half containing a decorative band did not captt 
more flying insects than the web half without 
decorative band (Tso, 1996). Similarly, A. keyserlv 
captured more prey on webs with decoration ban 
than on undecorated webs (Herberstein, 200 
Observations of A. appensa on Guam revealed 
differences    in    the    prey-interception    rates 
Silk decorations of orb-web spiders 663 
? 
decorated and undecorated webs, and no increase in 
prey interception in the decorated quadrants of the 
same webs (Hauber, 1998). This result was con- 
founded by differences in web size: decorated webs 
were significantly smaller than undecorated webs. 
After controlling for this size difference, decorated 
webs intercepted more prey per unit web area than 
undecorated webs (Hauber, 1998). 
The relationship between web decorations and 
prey capture has been examined in field studies of 
Cyclosa conica, which constructs linear decorations in 
its horizontal web (Tso, 19986) and Octonoba 
sybolides, which utilises linear and spiral decorations 
(Watanabe, 1999a). Decorated webs of G. conica 
intercepted 150% more insects than undecorated 
we-bSj even though they were 19% smaller than 
undecorated webs (Tso, 19986). Similarly, when the 
webs of 0. sybotides were adorned with either spiral or 
linear decorations, they intercepted more insects 
compared with undecorated webs. The size of 
decorated and undecorated webs did not differ 
(Watanabe, 1999a), although the mesh height of 
webs decorated with spirals was smaller than that of 
undecorated webs. However, this may not affect the 
conclusions because webs with a smaller mesh must 
also be more visible to insect prey than webs with a 
larger mesh (Craig, 1986; Watanabe, 1999a). 
Three manipulative experiments provide more 
conclusive evidence in support of a foraging function. 
In choice experiments, Drosophila spp. flies were 
confronted-with identical webs spun by Uhbprus 
glomosus (Craig & Bernard, 1990) and -O. sybotides 
(Watanabe, 1999 a). The web in one-arm of a Y- 
maze was illuminated with white light containing a 
UV component, while the web in the-other arm-was 
illuminated with white light without a UV com- 
ponent. Flies approached and were captured in webs 
.containing -U V-reflecting decorations more 
frequently than in webs that did not reflect UV. 
Furthermore, isolated bands of decorations built 
by A. aurantia were transferred onto an artificial web 
consisting of a synthetic adhesive mesh and exposed 
in the field (Tso, 1998a). In a control group, a 
similar-sized area of non-decorative silk was intro- 
duced into identical artificial webs. The artificial 
webs were then installed adjacent to already existing 
A. aurantia web sites. Artificial webs containing 
decoration bands captured significantly more flying 
insects than those in the control group, while web 
site or date had no effect on prey capture (Tso, 
1998 a). 
Blackledge and Wenzel (2000) recently published 
the results of an experimental study that tested the 
response of honeybees to silks spun by ancestral 
[Plerinochilus sp.) and derived spiders (A. aurantia) 
against a grass background. On the basis of their 
experiments they argue that they were unable to 
train bees to associate an award with UV-reflecting 
silk concluding that decorations spun by orb- 
spinning spiders are cryptic to bees which would not 
be able to discriminate UV-reflecting silk from 
background vegetation. 
Blackledge and Wenzel (2000) hypothesise that 
the reflectance spectra of the decorative silks.are flat 
but, regrettably, do not provide data to test this 
assertion. This is important because the only spectral; 
data available show a 30% variation- in the 
reflectance spectrum of decoration silks spun by A. 
argentata (-100% reflexive-at 370 nm varying to 
approximately-70% reflectance at 640 nm; Craig & 
Bernard, 1990). It is unknown whether there were 
any differences-.in the spectral properties of the 
ancestral and derived silks or the silks and their 
background. Further, most vegetation [with the ex- 
ception of densely hairy or glaucous leaves and 
foliage (Frolich, 1976)].absorbs, light in the UV to 
blue (330-420 nm) region of the spectrum and re- 
flects light in the green region of the spectrum and 
above ( > 550 nm). As a result, the contrast between 
the web decoration and a green background is high 
(see Figs 2 a andja in Craig & Bernard, 1990). 
Blackledge & Wenzel (2000) have shown that honey- 
bees had difficulty associating thtdecorative silks 
with the sugar reward, but learning is a different be- 
haviour from .perception. In fact, honeybees have 
more difficulty learning to associate a sugar water 
reward witfrUV-reflecting objects than objects that 
reflect any other wavelength (Menzel & Erber, 
1978). Finally, the-achromatic reflectance (bright-, 
ness) of the silks and the background was not- 
measured. Thus, we cannotdetermine if the observed 
choices were made in response to brightness or colour 
of the silks and their background. The experiments 
cited above show that understanding insect responses 
to colour, pattern and brightness is extremely com- 
plex and will require tightly controlled experiments 
that address one variable at a time. 
In a field experiment (Blackledge & Wenzel, 
1999), A. aurantia in decorated webs captured fewer 
prey than spiders in webs that had the decorations 
removed, suggesting that web decorations do not 
function to increase foraging success, but actually 
reduce it. This study is puzzling, since Tso (19986) 
used the same species (see above) and reported the 
opposite effect: artificial traps containing 
decorations intercepted more prey than control traps 
M E. Herberstein and others 
664 
? 
nossible explanations, blackleage <x v y 
5 id not control for web size in their study but To 
nWRb) used identical traps in both treatments. 
(19ja/>)   usea . b size between the Thus   random dinerences ui ?i.u > 
SSWSS'SSS and a? 
spiders ran Variation  in  the  attack 
Sedg   * ?   1999) may h?? difcen. 
Seville prey P?P^r(T?Pa.r998ir and 
aurmlia   in   Michigan,   USA   (iso,   IW? I 
iherefore serve a different function. 
Several other studies that investigate the source of 
varSSn in decorating hehavtour have ako proved 
A!n that are consistent with a foraging function. I or datathataec ed   (helr 
?5?| behaviour according to light condition, 
t mm light they adorned their webs with more and 
a
nrg
de?crtia,eydeeora,ions than in bng^hgh 
preLmahly   increasing  *? ???*; 
TJV reflection:  if decorations increase prey inter 
- SSS53S5wS* ??*? if web 
deeorTti"* are par. of a foraging strategy, spide s 
should vary the decorating behaviour ? ???? 
variation in the rate of prey encounter (?"?*"". 
Til   2000)   A. hymhni ?m maintained on the 
P
 rtlnt to note that a mechanical or anti-predator 
ZcZZoZpredict these changes in decorating 
IXlur according to ambient light condition or 
*??? response of ^-?? 
d^ti^^ 
decorations  (Blackledge,  1998ft).  In A. kyserhnp 
CaSjSn - o/.s 2000)  and .4.  trtffao*  d-* 
1999), food-deprived spiders also increased thesize 
of th   r webs. These patterns may reflect a trade-off 
be ween the costs and benefits of decorating be- 
haviour   In the field, decorated webs suffer trom 
'reacr   web   damage   due   to   prey   interception 
fHauber   1998). Consequently, hungry spiders may 
ieTceweb damage of their larger webs by spinning 
fewer decorations (Herberstein et al., 2000). I ^may 
also be that web decorations target ^-??? 
prev. In this case, starving spiders could be predl??V 
to spin large, undecorated webs that mtercept prey 
nom penally while well-fed spiders build sma ler 
decorated webs captunng finly those insects that 
^Sir^^Ssence of.decorating be 
havLm^amonglocturnal spiders, is ?*?jg 
with a foraeing function. Comparative analyses oe 
Twe n   he prelected by decorating and non- 
d^radng "specres   may   provide   further   insights 
into a foraging iunction. 
(4) Decreasing web damage 
The web decorations constructed by Argiope species 
or the silk tufts'spun by Gasteracantha^species are 
Conspicuous   structures, "and   may   advertise   the 
presence of webs to birds or other larger animals that 
PT otherwise- encounter  and  destroy   the  web 
Z ner &  Nowicki,  1983). This idea was tested 
xper mentally by placing paper replicates of cru- 
ciateTe orations made by A. fionda on the vacant 
^D^ul^ie.ofngturnaJv^W 
widens   During the day, webs with a paper strip 
Offered  a ^r^C^^n***^ 
.domed   webs   (Eisner   &   Nowicki,   1983).   It  B 
Scuhto aSSesl these data because the^sou? rf 
web destruction is not identified and there are a 
dumber of potentially biasing effects (see also Craig 
ft Semard 1990). For example, there was no control 
L dXnce! 1 the ability of webs to absorb insect 
kineu! energy,   the  primary   factor  causing  web 
hreikdown and which varies greatly between 
webs ofTfferent species (Craig, 1987) Henoe,som 
webs are destroyed when they mtercept large oi fas 
flvL  insects,   while  others  are  unaffected^  It  is 
un ikel    hat  he kinetic properties of the webs spun 
by sixdifferent araneid species are identical, and 
Aerefore the differences in damage may be caused 
bv varyL rates of insect interception. Similarly the 
reflec iv "properties of the pieces of paper used in 
he    rtudy we- not measured and it is likely that 
Irey offered from those of silk decorations, which 
Silk decorations of orb-web spiders 
may have very different effects on insect prey or even 
birds. 
However, a recent field experiment using the webs 
and decorations of A. aurantia showed that webs with 
decorations suffered less damage than webs that had 
all  decorations   removed   (Blackledge   &   Wenzel, 
1999), while the presence of the spider in the web did 
not affect the rate of web damage. Empty frames 
were initially set up in a triangular arrangement on 
a mown lawn and birds were lured to the middle of 
the arena with a dish of bird seed, After a period of 
acclimatisation, two of the three empty frames were 
replaced by a web containing decorations and a web 
where all  decorations had  been  removed.  Whi e 
Bhckledge   &   Wenzel   (1999)   did   not   provide 
information on .the. source of web destruction, they 
observed  several   birds   actively   avoid   webs   that 
contained decorations. 
In a different approach, Kerr (1993) related the 
frequency  of web-decorating  behaviour  with  the 
presence of birds on different Pacific islands, Only 
16 4?/  of A. appensa spun decorated webs on Guam, 
where  the  introduction  of the brown  tree snake 
(Boiga irregulans) has eliminated all of the native 
birds over the past 30 years (but see Hauber, 1998). 
By contrast, between 41.9 % and 56.9 % of A. appensa 
on the neighbouring islands of Rota, Tmian and 
Saipan*   where   the   bird   fauna   remains   intact, 
decorated  their  webs   (Kerr,   1993).   Perhaps  the 
extinction  of birds  on  these  islands  reduced  or 
eliminated selection for web-decorating behaviour. 
This evolutionary scenario suggests that the cost of 
decorating is very high if this behaviour is to be lost 
in such a short time interval. 
The main argument against an advertisement 
function is that many spiders that build decorations 
locate their webs in sheltered positions (see Table 2; 
Eberhard, 1990) such as tree buttresses (e.g. Lubinella 
spp.), tall grass (e.g. Argiope spp.) shrubs and 
undefstory (e.g. Gasteracantha spp.) where birds are 
unlikely to fly through and damage the web 
(Eberhard, 1990). Experimental designs that expose 
these webs outside their natural shelter (e.g. on a 
mown lawn in Blackledge & Wenzel, 1999) may not 
provide biologically relevant information. 
(5) Thermoregulation 
A few studies suggest a thermoregulatory function 
for disc-shaped decorations. Humphreys (1992) 
found that the discs constructed by juvenile Neogea 
spp. provided shade to spiders foraging in sites 
characterised by high temperatures. When tempera- 
tures exceeded 40 ?C, the spider moved under the 
disc where it was protected. The disc cut the 
transmission of light by 60%, reducing the body 
temperature of the spider by 1.8 ?C (Humphreys, 
1992). It is likely that this shuttling behaviour 
evolved secondarily, with the animal taking ad- 
vantage of an existing structure. However, web 
decorations are found in a variety of different light 
conditions, and more frequently under dim light 
(Elgar et al., 1996). Furthermore, linear or cruciate 
decorations do not shade the spiders' body and are 
thus of limited thermoregulatory benefit. 
VII. NON-FUNCTIONAL EXPLANATIONS 
Nentwig & Heimer '(1987) argue that the high 
degree of unexplained variation in decorating be- 
haviour, together with a body of contradictory 
evidence, indicates that web decorations are unlikely 
to have an evolutionary function. Instead, web- 
decorating behaviour may simply arise as a conse- 
quence of non-specific stress" reactions (Nentwig & 
Rogg, 1988) or silk regulation (Peters, 1993). 
Under laboratory conditions, A. argentata varied 
their decorating "behaviour in response to extreme 
temperatures, the presence -of males, moulting and 
age, factors that were thought to raise levels of stress 
(Nentwig & Rogg, 1988). However, these patterns 
were not replicated under field conditions, where 
variation  in   temperature,   humidity,  illumination 
and wind velocity failed to explain differences in 
web-decorating behaviour (Nentwig & Rogg, 1988). 
Peters (1993) suggests that web decorations are a by- 
product of silk regulation. Not all of the silk from the 
aciniform  (Araneidae)  and piriform  (Uloboridae) 
glands, is  used  in  wrapping  prey,  and   thus  the 
superfluous silk is transferred into the decorations 
(Peters,   1993).   This   idea   has   not   been- tested 
empirically. 
There are several objections to the non-functional 
explanation of silk decorations. First, the energetic 
cost of silk production is high (Peakall & Witt, 1976; 
Opell, 1998). Therefore, it is unlikely that natural 
selection will maintain web-decorating behaviour 
unless it provides some benefit. Furthermore, it 
would be surprising if a non-functional trait evolved 
independently several times, as appears to be the 
case. 
The stress-response interpretation is counter- 
intuitive; spiders that experience stress should not 
deplete their nutritional resources further by pro- 
ducing   more   costly   silk.   In   fact,   spiders   spin 
? ',-'   ' 
666 
decorations   under   extremely   benign   laboratory 
conditions that minimise or eliminate 'stress' (e.g. 
Herberstein et al., 2000). The inter-specific variation 
in web-decorating behaviour provides little support 
for   either   the   stress-response   or   silk-regulation 
explanations. For example, such explanations should 
apply equally to spiders that build webs during the 
day or night, but web decorating is not found in 
nocturnal species. Additionally, it is not clear why 
spiders belonging to the genus Argiope suffer higher 
stress levels or need to regulate silk more than spiders 
of, say, the genus Araneus that do not decorate their 
webs  '(Table   1).   Finally,   web   decorations   are 
typically constructed immediately after the orb web 
is completed.  Even if the spiders have produced 
superfluous silk, it seems unlikely that they would 
discard it at the beginning of a foraging bout when 
an unknown quantity of silk is needed to wrap prey. 
While an excess of silk probably does not induce 
decorating behaviour, it is likely that the availability 
of silk in the glands will affect the frequency and size 
of web decorations. The amount of wrapping silk 
expended   by   A.   aetherea- has   been   manipulated 
experimentally by offering some spiders more prey, 
whiclf was consequently wrapped but not ingested 
(I.M/Tso, unpublished data). These spiders spun 
fewer or smaller decorations in subsequent webs 
"than.spiders whose silk reserves were not depleted. 
At the end of the foraging period, the spider's reserve 
oraciniform silk was ..never completely exhausted, 
and spiders were always able to release more silk to 
wrap-prey (I. M. Tso, unpublished data). 
VIII. CONCLUSION 
The function of web decorations remains an un- 
resolved issue.  At least seven different functional 
explanations have been proposed, but only a few 
functions have been tested directly. The foraging 
function has been tested most extensively on linear, 
cruciate and circular decorations and is supported 
by several experiments and correlative field studies 
(e.g. Craig & Bernard, 1990; Craig, 1991; Hauber, 
1998- Tso, 1996, 1998 a, b; Watanabe, 1999 a). The 
anti-predatory function (Horton, 1980) and the web 
advertisement function (Blackledge & Wenzel, 1999) 
have   been   investigated   for   linear,   cruciate   and 
circular decorations built by Argiope spp. By contrast, 
the other functions and other types of decorations 
have not been tested as extensively. 
Assigning a specific function to web decorations is 
made difficult for a number of reasons. First, the 
M. E. Herberstein and others 
selective pressures responsible for the evolution of 
web decorations may differ from those identified by 
contemporary studies as maintaining the behaviour. 
Second, the convergent nature of decorations and 
their   patterns   implies    that   different    types   of 
decorations  serve  different  functions   in  different 
species. For example, the relationship between prey 
capture and web decorations is confirmed for the 
linear and cruciate types, particularly in Argiope spp. 
and may also apply to disc-shaped decorations spun 
by   some   Uloboridae   (Craig   &   Bernard,   1990; 
Watanabe, 1999 a). However, it is not clear whether 
debris types have similar functions. In fact, it is not 
clear whether  the  eggsacs placed Jn_the  web of 
Cyclosa, Arachnura and Gyrtophora species should be 
considered as decorations equivalent to the linear, 
circular or cruciate patterns. Keeping eggsacs in the 
web may allow the spider to protect the eggs from 
predators, parasites and parasitoids (see Elgar et al, 
1983). While these egg sacs-may provide additional 
concealment, the selective 'route' is.fundamentally 
different  to  those species that use  acinifo/m silk 
bands. Similarly, the prey items placed in bands in 
the webs of Nephila species may function primarily as 
food storage, with concealment being a selectively 
neutral consequence. ., 
Finally, the different types of decorations con- 
structed by juvenile and adult spiders also impede an 
.attempt to assign particular functions. While on- 
togenetic changes suggest distinct mechanisms or 
perhaps different functions for the different 
decorations types, these are still poorly understood. 
The vast majority of studies have been concerned 
with adult rather than juvenile spiders. 
Several steps may help to resolve the ongoing 
controversy    regarding     the    function    _qf    web 
decorations. First, it may be helpful to identify four 
phylogenetic clusters of web decorations:' uloborine' 
uloborids,     'argiopine'     and     'gasteracanthine' 
araneids and 'nephiline' tetragnathids. Within these 
clusters, similar web decorations may have similar 
functions as a result of common ancestry. Extra- 
polations from one phylogenetic locus to another are 
unlikely  to  be  relevant  in  resolving  this debate. 
Second,   the   different   decorating   patterns   may 
describe different phenomena that have undergone 
different selective routes. Therefore, the functions of 
the various types of decorations may or may not be 
convergent. For example, debris structures should 
not be termed 'decorations' or 'stabilimenta', but 
should be treated as separate behavioural phenom- 
ena. Two main clusters remain within the 'true' silk 
types of decorations. The 'bright white' silk bands 
Silk decorations of orb-web spiders 
spun frequently by species in the 'argiopine' and 
'uloborine' cluster, which may be convergent in 
form and function, (although evidence is not very 
strong at this stage), and the more dull decorations 
that are spun less frequently by species in the 
'gasteracanthine' and 'nepheline' clusters. 
Finally, inter-specific comparative studies that 
combine field observations, experimental studies and 
the life histories of decorating and non-decorating 
spiders should give the clearest insight into the 
selective factors that have influenced the evolution ol 
decorating behaviour among spiders. 
IX. ACKNOWLEDGMENTS 
We thank Astrid Heil.ng, I-Min Tso, Mark Hauber Bill 
Piel   and   two   anonymous   reviewers   for   their  helpful 
comments  and  criticisms.   Alexander  Kerr  generously 
provided a list of decorating species. Financial support 
was provided by the Austrian Science Foundation (J150Q- 
BIO)   to   M.E.H.,   the   Australian   Research   Council 
(A19930103) l0 M.A.E., the Bunting Institute and the 
American Association of University Women (AAUW) to 
C  1   C and NSF (NSF DEB-9712353; DEB-97-07744), 
the Smithsonian-Neotropical  Lowlands  and  Scholarly 
Studies Programs, to J. A. C. 
X. REFERENCES 
ADAI'0-BARR.ON, A. M. & BARRION, A. T. (1981), Spider prey 
of   Sceliphrm   madraspatanum    conspicillatus    (Hymcnoptera: 
Sphecidae}-. Kalikasan 10, 122-125. 
AUSTIN, A. D. & ANDERSON, D. T. (1978). Reproduction and 
development of the spider Nephila edulis (Koch) (Araneidae: . 
Araneae). Ausl. J. <W. 26, 501-518. 
BLACKLEDGE,  T. A.   (1998a).  Signal  conflict ia spider  webs 
driven by predators and prey. Proc. R. Soc. Land. B. 265, 
1991-1995. 
BLACKLEDGE,   T. A.   (19984).   Stabilimentum   variation   and 
foraging  success  in  Argiope aurantia  and  Argiope  trifasaata 
(Araneae, Araneidae). J. Z??l- Lond. 246, 21-27. 
BLACKLEDGE, T. A. &WENZEL,J. W. (1999). Dostabilimentain 
orb webs attract prey or defend spiders? Behav. Ecol.  10, 
372-376. 
BLACKLEDGE, T. A. & WENZEL, J. W. (2000) The evolution of 
cryptic spider silk: a behavioral test. Behav. Ecol. 11, 142-145. 
BRISTOWE, W. S. (1941). A comity of spiders. The Ray Society, 
London. 
CLOUDSLEY-THOMPSON, J. L.   (1995).  A  review  of the  anti- 
predator devices of spiders. Bull. Br. Arachnol. Soc. 10, 81-96. 
CLYNE,  D.   (1969).  A guide  to Australian  spiders.  Thomas 
Nelson, Ltd. Australia. 
CODDINGTON, J. A. (1990). Ontogeny and homology in the male 
palpus   of orb-weaving   spiders   and   their   relatives,   with 
comments on phylogeny (Araneoclada: Araneoidea, Deino- 
poidca). Smithsonian Contrib. Z?oL 496> '~52- 
667 
COMSTOCK, J. H. (1912).  The spider book. Doubleday, Page & 
Company, New York. 
COMSTOCK,.]. H. (1940). The spider book. Doubleday, Doran & 
Company, Inc., New York. 
CRAIG C L  (1986). Orb visibility, the influence of insect Hignt 
J
 behaviour and visual physiology on the evolution of web 
designs within the Araneoidea. Amm. Behav. 34, 54-68. 
CRAIG   C L.   (1987).  The significance of spider size  to the 
diversification  of spider-web   architectures  and   spider   re- 
productive modes. Am. Nat. 129, 47-68. 
CRAIG, C. L.  (1989). Alternative foraging modes of orb web 
weaving spiders. Biotropica 21, 257-264. 
CRAIG C L  (1991). Physical constraints on group foraging and 
social evolution: observations on web-spinning spiders. Fund. 
Ecol. 5, 649-654. 
CRAIG   C L   (1994a). Limits to learning: effects ol predator 
pattern and colour on perception and avoidance-learning by 
prey. Anim. Behav. 47, 1087-1099. 
CRAIG C L (1994A). Predator foraging behavior in responses 
' perception and learning by its prey: interactions between orb- 
spinning spiders and stingless bees. Behav. Ecol. Soaobwl. 35, 
45-52 
CRAIG, C. L. (1997). Evolution of arthropod silks. Annu. Rev. 
-.Entomoi. 42, 231-267.   
CRAIG   C L. & BERNARD, G. D. (1990). Insect attraction to 
ultraviolet-reflecting spider webs and web decorations. Eulogy 
71.616-623. I   A    /,QQi\ 
- CRAIG, C. L., BERNARD, G. D.& CODDINGTON, J. A; (U )? 
Evolutionary shifts in the spectral properties of spider silks. 
Evolution 48, 287-296. 
CUSHING, P. E. & OPELL, B. D. (1989). Disturbance behaviours 
in the spider Uloborus glomosus (Araneae, Ulobndae). Can. J. 
Zool. 68, 1090-1097. ^ 
EBERHARD, W. G. (1973). Stabilimenta on the webs of Uloborus 
diversus  (Araneae: Uloboridae)  and other spiders. J. Zool. 
Lond. 17LJ67-384. 
EBERHARD, W. G. (1990). Function and phylogeny of spider 
webs". Armtr.Rev.Ec'ol. Syst. 21, 341-37E    : 
EDMUNDS   J   (1986). The stabilimenta of Argiope flampalpis and 
Areiope 'tnfasciala in West Africa, with a discussion of the .. 
function of stabilimenta. In Proceedings of the 9th International 
Confess of Arachnology, Panama 1983 (eds-W. G, Eberhard, 
Y.D.Lubin and B. C. Robinson.), pp. 61-72. Smithsonian 
Institute Press, Washington. 
EDMUNDS, J. & EDMUNDS, M. (1986). The defensive mechanisms 
of orb weavers (Araneae: Araneidae) in Ghana, West Africa. 
In Proceedings of the 9th International Congress of Arachnology, 
Panama 1983 (eds W. G. Eberhard, Y. D, Lubin and B.C. 
Robinson.),    pp.    73 89.    Smithsonian    Institute    Press, 
Washington. 
EISNER   T    &   NOWICKI,   S.   (1983).   Spider  web  protection 
through  visual   advertisement:   role  of the  stabilimentum. 
Science 219. 185-187. 
ELGAR  M. A., ALLAN, R. A. & EVANS, T. A. (1996). Foraging 
strategies  in  orb-spinning spiders:   ambient  light  and  silk 
decorations    in    Argiope    aetherea    Walckenaer    (Araneae: 
Araneoidea). Aust. J. Ecol. 21, 464-467. 
ELGAR M A. &JEBB, M. (1999). Nest provisioning in the mud- 
dauber wasp Sceliphron laetum (F. Smith): body mass and taxa 
specific prey selection. Behaviour 136, 147-159. 
ELGAR M A., POPE, B. & WILLIAMSON, I. (1983). Observations 
on the spatial distribution and natural history of Cyrtophora 
668 
M. E. Herberstein and others 
hirta    (L. Koch)    (Araneae:    Araneidae)    in    Queensland, 
Australia. Bull. Br. Arachnol. Soc. 6, 83-87. 
EWER, R. F. (1972). The devices in the web of the African spider 
Argiope flavipalpis. J. Mat. Hist. 6, 159-167. 
FOELIX, R. F. (1992). Biologic der Spinnen. Gcorg Thieme Verlag 
Stuttgart. 
FROLICH, M. W. (1976). Appearance of vegetation in ultraviolet 
light: absorbing flowers, reflecting backgrounds  Science 194 
839-841. 
GRISWOLD, C. E, CODDINGTON J. A., HORMIGA, G., & SCHARFF, 
N.   (1998).   Phylogeny   of  the   orb-web   building   spiders 
(Araneae, Orbiculariae: Deinopoidea, Araneoidea)   Zool   7 
Linn. Soc. 123, 1-99. 
GRISWOLD,   C. E.,   CODDINGTON, J. A.,   PLATNICK,   N I     & 
FORSTER, R. R. (1999). Towards a phylogeny of entelegyne 
spiders (Araneae, Entelegynae). J. Arachnol. 27  53-63 
HAUBER,   M.E.   (1998).   Web   decorations   and   alternative 
foraging tactics of the spider Argiope appensa. Elhol. Ecol Evol 
10, 47-54. 
HERBERSTEIN, M. E.  (2000). Foraging behaviour in orb-web 
spiders (Araneidae): do web decorations increase prey capture 
success in Argiope keyserlingi Karsch 1978? Aust. 7 Zool  48 
217-223. '     ' 
HERBERSTEIN, M. E.,  CRAIG, C. L.  &  ELGAR,  M. E.  (2000). 
Foraging strategies and feeding regimes: web and decoration 
investment in Argiope keyserlingi Karsch Mraneae: Araneidae) 
Evol. Ecol. Res. 2, 69-80. 
HINGSTON, R.-W. G. (1927). Protectivedevices in spider's snares, 
with a description of seven new species of orb-weaving spiders 
Proc. Zool. Soc. Lond. 28, 259-293. 
HORMIGA, G., EBERHARD, W. G. & CODDINGTON, J. A. (1995). 
Web construction behavior in Australian Phonognatha and the 
phylogeny of nephiline ami tetragnathid spiders (Araneae, 
Tetragnathidae). Aust. J. Zool. 43, 313-343. 
HORTON, C. C. (1980). A defensive function for the stabilimenta 
of two orb weaving spiders (Araneae: Araneidae). Psyche 87 
13-20. -_ ----- 
'. HUMPHREYS, W. F. (1992). .Stabilimenta as parasols: shade 
construction by Meogea sp. (Araneae: Araneidae, Argiopinae) 
and its thermal behaviour,.Bull. Br. Arachnol. Soc. 9, 47-52. 
KERR, A. M. (1993). Low frequency of stabilimenta in orb webs 
of Argiope appensa (Araneae: Araneidae) from Guam: an 
indirect effect of an introduced avian predator. Pacific Science 
47, 328-337. 
KOVOOR, J. (1987). Comparative structure and histochemistry 
of silk-producing organs in arachnids. In Ecophysiology of 
spiders (ed. W. Nentwig), pp. 160-186. Springer Verlag, 
Berlin. 
KULLMANN, E. (1958). Beobachtung des Netzbaues und Beitrage 
zur Biologie von Cyrtophora citricola Forskal (Araneae, 
Araneidae). Zool. Jb. Syst. 86, 181-216. 
LEVI, H. W. (1977). The orb-weaver genera Cyclosa, Melazygia 
and Euslala north of Mexico (Araneae: Araneidae). Bull. Mus 
Comp. Zool. 148, 61-127. 
LEVI, H. W. (1983). The orb-weaver genera Argiope, Gea, and 
Meogea from the western Pacific region (Araneae: Araneidae, 
Argiopinae). Bull. Mus. Comp. Zool. 150, 247-338. 
LEVI, H. W. (1986). The orb-weaver genus Witica (Araneae- 
Araneidae). Psyche 93, 35-46. 
LUBIN, Y. D. (1975). Stabilimenta and barrier webs in the orb 
webs of Argiope argentata (Araneae, Araneidae) on Daphne 
and Santa Cruz Islands, Galapados. J. Arachnol. 2, 119-126. 
LUBIN V. D. (1986). Web building and prey capture in the 
Ulobondae. In Spiders- Webs, behavior, and evolution (ed W A 
Shear), pp. 132-171. Stanford University Press, Stanford, 
C*A. 
LUBIN, Y. D., OPELL, B. D., EBERHARD, W. G. &LEVI, H. VV. 
(1982). Orb plus cone-webs in Uloboridae (Araneae),' with a 
description of a new genus and four new species. Psyche 89, 
MADDISON, VV. P. & MADDISON. D. R. (1992). MacClade- 
Interactive analysis of phylogeny and character evolution 
Version 3.0. Sunderland, IVlassachusetts: Sinauer Associates. 
MALT, S. (1993). Trophicai relations of selected web-building 
spiders lAraneae) in xerophil grasslands. Boll. Ace. Gioenia Sci 
Nat. 26, 253-261. 
MARPLES, B.J. (1969). Observations on decorated webs. Bull. 
Br. Arachnol. Soc. 1, 13-18. 
MARSON, j.E. (1947a). Some observations on the ecological 
variation and development of the cruciate zigzag camouflage 
device of Argiope pulchella (Thor.). Proc. Zool. Soc. Lond. 117 
219-227. 
MARSON, J. E. (1947A). Some observations on the variations in 
the camouflage devices used by Cyclosa insulana (Costa), an 
Asiatic spider, in its web. Proc. Zool. Soc. Lond. 117, 598-605 
MASCORD, R. (1970). Australian spiders in colour. Charles E 
Tuttle Co., Vermont. 
MCCOOK, H. C. (1889). American spiders and their spinningwork I 
Academy of Natural Sciences of Philadelphia, "Philadelphia 
MENZEL, R. & ERBER, J. (1978). Learning and memory in bees 
Scient. Am. 239, 102-109. 
NEET, C. R. (1990). Function and structural variability of the 
stabilimenta of Cyclosa insulana (Costa) (Araneae, Araneidae) 
Bull. Br. Arachnol. Soc. 8, 161-164. 
NENTWIG, W. & HEIMER, S. (1987). Ecological aspects 6tspider 
webs. In Ecophysiology of spiders (ed. W. Nentwig), pp. 211-225 
Springer Verlag, Berlin. 
NENTWIG, VV. & Rooo, H. (1988). The cross stabilimentum of 
Argiope argentata (Araneae: Araneidae) - nonfunctional or a 
nonspecific stress reaction? Zool. Ana. 221   246-266 - 
OPELL,  B. D.   (1979).   Revision  of the genera  and' tropical 
American species in the spider family Uloboridae. Bull Mus 
Comp. Zool. 148, 445-549. 
OPELL, B. D. (1987). The new species Philoponella herediae'&nd its 
modified orb-web  (Araneae," Uloboridae). J. Arachnol.  15, 
59-63. 
OPELL, B. D. (1998). The material cost and stickiness of capture 
threads and the evolution of orb-weaving spiders Biol T. 
Linn. Soc. 62, 443-458. ' 
PEAKALL, D. B. & WITT, P. N. (1976). The energy budget of an 
orb web-building spider. Comp. Biochem. Physiol 54A 187-190 
PEASLEE, J. E. & PECK, W. B. (1983). The biology of Octonoba 
octonanus (Muma)   (Araneae,  Uloboridae).   7. Arachnol   11 
51-67. *' 
PETERS, H. M. (1993). Uber das Problem der Stabilimente in 
Spinnennetzen. Zool. Jb. Physiol. 97, 245-264. 
RABAUD,   E.   (1932).   Le   role   du   stabilimentum   des   toiles 
d'Argiope fasciee. C. R. Acad. Sci. 194, 655-657. 
ROBINSON, B.C. & ROBINSON, M. H. (1974). The biology of 
some Argiope species from New Guinea: predatory behaviour 
and stabilimentum construction (Araneae: Araneidae)  Zool 
J. Linn. Soc. Lond. 54, 145-159. 
ROBINSON, B. & ROBINSON, M. H. (1978). Developmental studies 
of    Argiope    argentata     (Fabricius)     and     Argiope    aemula 
(Walckenaer). Symp. Zool. Soc. Lond. 42, 31^0. 
^ 
Silk decorations of orb-web spiders 
ROBINSON, M.H. & LUB.N, Y. (1979). Specialists and 
generalises: the ecology and behavior of some web-bu.ldmg 
spiders from Papua New Guinea. I. Herennia omatissima, 
Arg>op? ocyaloides and4r?fomm??to?*w(Arancae: Arancidae). 
Par./?. 21,97-132. 
ROB.NSON, M. H. & ROBINSON, B. (1970). The stabilimentum of 
the orb web spider, Argiope argenlata: an improbable defence 
against predators. Can. Enl. 102, 641-655. 
ROB.NSON,   M.H.   &   ROBINSON,   B.   (1973a).   Ecology   and 
behavior of the giant wood spider tophila maculala (* abacus) 
in New Guinea. Smith. Contr. fool. 149, 1-76. 
ROBINSON, M. H. & ROB.NSON, B. (19734). The stabilimenta of 
.Nephila clampes and the origins of stabilimcntum-bu.lding in 
araneids. Psyche 80, 277-288.    - 
ROBINSON, M.H. ScROB.NSON, B. (1980). Comparative slud.es 
of the courtship and mating behav.or of trop.cal arane.d 
spiders. Pac. Ins. Mono. 36, 1-218. 
ROVNER   J  S    (1976).  Detritus stabilimenta on  the webs ot 
Cydosa turbinata (Araneae, Araneidae). J. Arachnol. 4,215-216. 
SCHARFF    N.   &   CODDINOTON, J.A.   (1997).   A   phylogent.c 
analysis    of   the    orb-weaving   spider   farmly    Araneidae 
(Arachnida, Araneae). Zool. J. Linn. Soc. 120, 355-424. 
SCHOENER,   T.W.   &   SP.LLER,   D. A.   (1992).   Stabilimenta 
characteristics of the spider Argiope argenlata on small islands: 
support   of  the   predator-defense   hypothesis.   Behav.   Ecol. 
Sociobiol. 31, 309-318. 
SIMON, E. (1895). Historie naturelle des Araignees. Roset, Pans. 
669 
THIRUNAVUKARASU, P., NICOLSON, M. & ELGAR, M. A. (1996). 
Leaf selection by the leaf-curling spider Phonognatha graejei 
(Keyserling) (Araneoidea: Araneae). Bull. Br. Arachnol. Soc. 
10, 187-189. 
TOLBERT.W. W. (1975). Predator avoidance behaviors and web 
defensive structures in the orb weavers Argiope aurantia and 
Argiope trifasciala (Araneae: Araneidae). Psyche 92, 29-52. 
Tso   I. M. (1996). Stabilimentum of the garden spider Argiope 
trifasciala:* possible prey attractant. Anim. Behav. 52, 183-191. 
Tso  I. M. (1998a). Isolated spider web stabilimentum attracts 
insects. Behaviour 135, 311-319. 
Tso   I  M.  (1998A). Stabilimentum-decorated webs spun by 
Cydosa conica (Araneae, Araneidae) trapped more insects than 
underrated webs. J. Arachnol 26, 101-105. 
Tso   I  M   (1999). Behavioral response of Argiope trifasciala to 
recent foraging gain: a manipulative study. Am. Midi. Nat. 
141,238-246. 
WATANABE, T. (1999a). Prey-attraction as a possible function ot 
the silk decoration of the uloborid spider Oclonoba sybohdes. 
Behav. Ecol. 5, 607-611. 
WATANABE, T. (1999A). The influence of energetic state on the 
form of stabilimentum built by Oclonoba sybohdes (Araneae: 
Uloboridae). Ethology 105, 719-725. 
WIEHLE H   (1929). Weitere Beitrage zur Biologie der Araneen, 
insbesondere zur Kenntnis des Radnetzbaues. Z Morph. Okol. 
Tiere 15, 262-306. 
YAGINUMA, T.  (1960). Spiders of Japan in colour. Hoikusae, 
Osaka.