REGIONAL SEISMIC SONG DIFFERENCES IN SKY ISLAND POPULATIONS OF THE JUMPING SPIDER HABRONATTUS PUGILLIS GRISWOLD (ARANEAE, SALTICIDAE)

2006·14.1 MB·English

by  Damian O Elias

2006 (2007). The Journal of Arachnology 34:545-556
REGIONAL SEISMIC SONG DIFFERENCES IN SKY ISLAND
POPULATIONS OF THE JUMPING SPIDER
HABRONATTUS PUGILLIS GRISWOLD
(ARANEAE, SALTICIDAE)
Damian O. Elias, Eileen A. Hebets,^ Ronald R* Hoy5"^ Wayne P. Maddison^ and
Andrew C. Mason*: “Division of Life Sciences, Integrative Behaviour and
Neuroscience, University of Toronto at Scarborough, 1265 Military Trail,
Scarborough, Ontario, MIC 1A4, Canada; ^Departments of Zoology and Botany,
University of British Columbia, Vancouver, British Columbia, V6T 1Z4, Canada;
^School of Biological Sciences, University of Nebraska, Lincoln, Nebraska 68588,
USA; "“Department of Neurobiology and Behavior, Cornell University, Ithaca, New
York 14850, USA.
ABSTRACT. Jumping spiders have long been used as model organisms to study visual communication.
However, recent studies documenting thepresence ofintricate multicomponent seismic songs duringcourt-
ship displays suggest an important role of seismic communication as well. Given the relatively recent
focus on seismic communication, the extent to which seismic songs vary among jumping spider species
or even among populations remains poorly understood. Here, we use the extensively studied Habronattus
pugillis Griswold 1987 complex to explore putative seismic song diversity among males from isolated
populations. H. pugillis populations have been studied extensively because of the tremendous diversifi-
cation of male visual secondary sexual ornaments observed among adjacent mountain-top populations in
southeastern Arizona (“sky islands”). Here, we aim to explore putative parallel patterns ofdiversification
in seismic courtship songs between different sky island populations. Using laser vibrometry, we examined
in detail the songs of three mountaintop populations (Atascosa (AT), Santa Rita (SR), and Santa Catalina
(SC)) and observed an extraordinary diversity of songs and song types among these three populations.
Largedifferences were seen in both thetemporal and spectral properties ofmale seismic songs. Inaddition,
we observed differences in song complexity between populations with some populations having “simple”
songs (1 component) and others having “complex” songs (3 components). We also present preliminary
data from four additional populations (Galiuro (GA), Huachuca (HU), Mule (MU), and Patagonia (PA)).
Results from this study suggest that the diversification ofmale visual signals observed among populations
ofH. pugillis finds a counterpart in male seismic songs
Keywords: Seismic communication, signal evolution, species diversification, Salticidae
Jumping spiders have proven to be fruitful Forster 1982a,1982b; Richman 1982; Clark &
&
models in the study of ecology, behavior, and Uetz 1990, 1992, 1993; Tarsitano Jackson
evolution, particularly as it relates to visually 1992, 1994, 1997; Edwards & Jackson 1994;
&
guided behaviors (Land 1969a, 1969b, 1985; Jackson Pollard 1996; Harland et al. 1999;
Eakin & Brandenburger 1971; DeVoe 1975; Harland & Jackson 2000, 2001, 2002; Naka-
Williams & McIntyre 1980; Blest et al. 1981; mura & Yamashita 2000; Taylor et al. 2000,
Land & Nilsson 2002). Studies have demon- 2001; Clark & Morjan 2001; Jackson et al.
strated the amazing visual abilities thatjump- 2005; Li & Lim 2005; Nelson et al. 2005;
& &
ing spiders possess by focusing on behaviors Hoefler Jakob 2006; Nelson Jackson
&
from a variety ofcontexts including predatory, 2006; Su Li 2006). However, recent re-
navigational, mating, and competitive inter- search has highlighted the utilization of seis-
actions (Crane 1949; Jackson 1977; Hill 1979; mic (vibratory) songs during courtship dis-
plays (Jackson 1977, 1982; Edwards 1981;
5Corresponding author. E-mail: elias@utsc. Gwynne & Dadour 1985; Maddison & Strat-
utoronto.ca ton 1988a, 1988b; Noordam 2002; Elias et al.
545
546 THE JOURNAL OF ARACHNOLOGY
2003) and a few studies have demonstrated tain range, males are very similar (Maddison
that these seismic songs are crucial for mating & McMahon 2000). Using a combination of
success (Elias et al» 2004, 2005, 2006a). De- behavioral, molecular, and phylogenetic data,
spite the recent increase in studies focused on Masta & Maddisoe (2002) demonstrated that
seismic communication in jumping spiders, sexual selection was driving the observed di-
we still know very little about the taxonomic versification of male traits. Hebets & Maddi-
breadth and/or importance of this mode of son (2005) then suggested that a process of
communication within the family Salticidae. antagonistic co-evolution (Holland & Rice
Jumping spiders in the genus Habronattus 1998) couldbe responsible fordriving among-
have been the subject of extensive studies fo- population variation in female mating prefer-
cused on species diversification, phylogeog- ences and associated male traits. In a recip-
raphy, communication, mate choice, signal de- rocal mate choice study, they found a
sign, and sexual selection (Griswold 1987; H
Cutler 1988; Maddisoe & Stratton 1988a, xeeophilic mating preference in which
1988b; Masta 2000; Maddison & McMahon pugillis females from the Santa Rita Moun-
2000; Masta & Maddison 2002; Elias et aL tains preferred males from a foreign popula-
H20e0d3i,n22000043,;20H0e5b,et2s00&6a,Ma2d0d0i6sbo;eMa2d00d5i)s.onNo&t toiwonn, lforcoaml tmhaeleAsta(sHceobseatsMo&untMaaidndsi,sooveer20t0h5e)i.r
Male H, pugillis from the Atascosa (AT) and
only is this genus diverse, with over 100 spe-
cies described in North America (Griswold Santa Rita (SR) populations were also recent-
1987; Maddison & Hedin 2003), but it also ly observed to produce complex seismic songs
incorporates extensive morphological and be- (Elias et al. 2005, 2006a). In afollow-up study
havioral differentiation among its many spe- exploring the previously established SR fe-
cies. Habronattus males include some of the male preference forATmales (Hebets & Mad-
most elaborate male ornamentation and visual dison 2005), Elias et al. (2006a) demonstrated
courtship behaviors known among any spider that a female bias for complex/novel seismic
species (Peckham & Peckham 1889, 1890; signals was responsible for the observed xee-
Griswold 1987; Maddison & Hedin 2003). In ophilic preference and suggested that a gen-
addition to their elaborate ornamentation, it eral bias for complexity/novelty among fe-
was recently demonstrated that some species males could have contributed to the rapid
of Habronattus produce complex m.ulticom- diversification observed in the H. pugillis
ponent seismic songs (vibrations) simulta- group (Masta 2000; Masta & Maddison
neous with visual signals during courtship 2002
& ).
(Maddisoe Stratton 1988b; Elias et aL The goal ofthis particular study was to doc-
2003, 2005, 2006a). Furthermore, these seis- ument and compare seismic courtship songs
mic signals were shown to be a crucial factor of H. pugillis populations. In so doing, we
in mating decisions (Elias et aL 2004, 2005). demonstrate that the striking diversity of vi-
One particularly well studied group of Ha- sual displays observed among populations of
bronattus are those in the H. pugillis complex.
H. pugillis finds a counterpart in seismic
In North America, populations of H. pugillis songs among populations. We detected differ-
Griswold 1987 occur in woodland habitats
ences in both the temporal and spectral prop-
isolated at the top of mountain ranges in
erties of songs between populations. In addi-
southeastern Arizona and into Mexico. These
mountain ranges are known as “sky islands” tion, we observed differences in the
complexity of songs, with some populations
because their peaks form an archipelago of
isolated woodlands separated by desert low- having songs with a single component and
lands (Warshall 1995). Males from these var- others having multiple components. This is
ious isolated populations are exceptional in the first study to our knowledge to show re-
that each possesses distinct secondary sexual gional differences in spider songs. We suggest
traits involving both morphological and be- that seismic songs along with visual orna-
havioral differences (Maddison & McMahon ments are under strong selection in H. pugillis
2000; Elias et aL 2006b). While there is an and we discuss the evolutionary forces that
impressive amoog-population variation in H. may have driven this diversification ofseismic
pugillis males, within a population or moun- songs.
—
ELIAS ET AL.—SEISMIC SONGS IN HABRONATTUS PUGILLIS 547
METHODS and thus we have no data on male age. We
—
recorded seismic vibrations using a laser dop-
Spiders. Male and female H. pugillis were pler vibrometer (LDV) (Polytec OFV 3001
collected from different mountain ranges in Ar- controller, OFV 511 sensor head) (Michelsen
izona (Atascosas (AT) 3r24.63'N, 111°8J7'W;
Santa Ritas (SR) 3r4038'N, 110°52.82'W; et al. 1982). Pieces of reflective tape (approx.
Santa Catalinas (SC), 32°2L40'N, 110°55.37'W; 1 mm^) were attachedmtmo the underside of the
Hm courtship substrate 2 from the female to
Galiuros (GA), 32°3438'N, 110°16.50'W;
serve as measurement points forthe LDV. The
achucas (HU) 3r25.94'N, 110°17.50'W; Pata- LDV signal was recorded on the audio track
gonias (PA) 3r23.87'N, 110°44.44'W, and
during standard video taping of courtship be-
Mules (MU) 31°29.68'N, 109°59.82'W) over havior (Sony DVCAM DSR-20 digital VCR,
three field seasons (April-June 2002, April- 48 kHz audio sampling rate). Spectrograms
May 2003, April-May 2004). Males and fe- were made using Raven software (Cornell
males were collected as immatures and adults. University, Lab of Ornithology). We present
Male courtship songs were recorded up to a
detailed measurements of spider songs of
maximum of 3 mo after the animals were col- = —
three populations (AT, n 15; SR, n 12;
lected. As males senesce, they cease to initiate SC, n = 12). Means are given ± SD.
courtship and instead avoid or act aggressive to- As it was not possible to record songs for
wards females (Elias, pers. obs.), thus only all the populations using LDV, we presentpre-
males that actively courted females were used. liminary data on songs from four more pop-
cAonnitmaailnesrswe(rAeMAhoCusPeldastiincdiPvriodduuacltlsy, Pinetapllausmtai,c u3;laHtiUo,nsn(G=A,3)nth=at1w0e; PrAe,cornde=d 5u;siMngU,a ncus—-
CA; 3 X 3 X 5cm) and kept segregated by sex. tom piezoelectric sensor built from a turntable
Animals were kept in the lab on a 12:12 light: needle cartridge. For this recording technique,
dark cycle. Spiders were fed fruit flies {Dro~ the courtship arena was a sheet ofgraph paper
sophila melanogaster) and juvenile crickets attached to a square cardboard frame (60 X
(Acheta domesticus) once aweek. Male voucher 45 cm). Females were tethered as above and
specimens are deposited at the Royal Ontario the male’s seismic signals were recordedusing
Museum, Toronto, Canada.
a piezo-electric sensor placed directly under-
Recording procedures and analysis. neath the tethered female. Ten different fe-
Detailed measurements on seismic songs were males were used in piezo-electric recordings.
made using laser vibrometry (Elias et al. In comparing populations where we recorded
2003). We first anesthetized a mature female a male’s signal using both LDV and the piezo-
H. pugillis with CO2 and tethered herto a wire electric sensor, we observed that although low
with low melting point wax (beeswax). We frequency responses (<150 Hz) were relative-
held females in place with a micromanipulator ly attenuated by the piezoelectric sensor, the
on a substrate of nylon fabric (25 X 30 cm) male’s signals were not significantly altered
stretched across a needlepoint frame to stan- and all signal components were apparent al-
dardize the tension of the nylon. As courting beit at lower amplitude (Elias et al. 2003), All
substrate has significant effects on signal piezo recordings were conducted in a sound-
transmission (Magal et al. 2000; Cokl et al. attenuated chamber at Cornell University.
2004, 2005; Elias et al. 2004), we used the Seismic signals were amplified (Nikko
nylon fabric as our courting surface since it NA790), recorded on the audio track ofa vid-
has negligible resonance characteristics and eo recording as above (48 kHz audio sampling
passes all frequencies equally (Elias et al. rate) and high-pass filtered (> 150 Hz), We
2003, 2006c), thus enabling us to observe all present examples oftypical spider songs from
the potential temporal and frequency compo- the recordings available.
nents of a male’s song. Mature males were As all recordings were conducted with teth-
dropped individually onto this substrate 15 cm ered females, it is possible that males behave
from the female and allowed to court freely. differently under these conditions than they
Recordings began when males orientated to- would in the field. These differences however
wards females. Fifteen different females were appear to be more in the duration ofcourtship
used to initiate courtship from thirty-nine ma- displays and not in the individual song com-
ture males. Males were collected as matures ponents. For example, in situations where fe-
548 THE JOURNAL OF ARACHNOLOGY
males were not tethered, males courted for Some populations only include the crackle
longer durations overall, but used the same component (see below) and these tend to have
song components (Elias, pers. obs.). In addi- crackles that are longer in duration than pop-
tion, since males were collected as matures, ulations with additional components. The sec-
they may have previously mated in the field. ond component (B) (“rasp”) is generally long
Here, also, we suggest that previous experi- in duration and broad in frequency. Rasps oc-
ence is unlikely to alter the specific compo- cur in all populations immediately prior to an
nents of seismic signal production but, in- attempted mount but, in some populations,
stead, alters more plastic behaviors such as from long distances. The third component (C)
courtship duration or latency to court. (“drone”) is short in duration and broad in
frequency but occurs at lower frequencies
RESULTS
than crackles. Drones occur in bouts consist-
Visual courtship signals in H, pugiltts,— ing of multiple signals produced rapidly fol-
The visual courtship behavior of H. pugillis lowing each other.
varies by population, but in general courtship All H. pugillis songs are composed of sim-
can be divided into two main stages: (1) the ilar song components, but there is variation in
approach stage and (2) the pre-mount stage (1) the types and number of components and
(Maddison & McMahon 2000). In the ap- (2) the temporal and spectral characteristics of
proach stage, the male raises and spreads his the different components. Below are detailed
first pair of legs and lowers and spreads his descriptions of male songs of three different
palps. The male then proceeds to approach the sky island populations followed by prelimi-
female either directly or in a sidling motion nary descriptions of four additional popula-
while flicking (rapidly moving) his forelegs tions. —
and pedipalps in a stereotyped manner. When Courtship behavior ofH,pugiliis, Santa
the male gets within one to two body lengths Rita (SR) males: The courtship behaviorofSR
of the female, the pre-mount stage begins males begins with rotations ofthe palps (Mad-
(Maddison & McMahon 2000), In the pre- dison & McMahon 2000), This palpal rotation
mount stage, the male’s approach slows down, is unique to SR males and is continued
and leg and pedipalp flicking becomes m.ore throughout the courtship display. Palpal rota-
—
rapid- especially downward flicks of the first tions are often punctuated with rapid leg
pair of legs. Males also scrape the abdomen flicks. Males remain mostly stationary during
repeatedly against the carapace just prior to courtship until the actual approach to the fe-
mounting a female. It was suggested by Mad- male, which is generally direct rather than si-
dison & McMahon (2000) that this grinding dling. The final stages ofcourtship involve the
corresponded to the production of seismic male holding his first pair of legs above the
songs. — female and flicking the tips. Leg flicking oc-
Seismic songs in H, pugillis^ Seismic curs less often than in other populations (i.e,,
songs are, in fact, produced by H, pugillis AT, SC, PA, HU). Leg flicks are coordinated
males at the moment the abdomen is seen to with seismic songs (Fig. 1) which consist of
rub against the carapace. Preventing the ab- a single component. SR seismic songs are of
domen from moving relative to the carapace variable duration but are generally short (0.56
prevents song production (Elias et al. 2006a). ± 0.393 s; « = 17) and consist of high inten-
Song production varies from population to sity, broad band (range: 0-“2850 Hz; peak fre-
population and males can produce songs in the quency: 1082 ± 540 Hz; n = 17) crackles
approach stage and/or the pre-mount stage. (labelled “a” in Fig, 1). Some SR males in-
Song production is usually coordinated with clude rasps at extremely shortrangesjustprior
flicking of forelegs. The general H, pugillis to attempted copulation (Fig. 1, 48-55 s, SR
song can have three components. The first column). The majority ofseismic signals how-
component (A) (“crackle”) is generally of ever only include the crackle component (Fig.
short duration, broad frequency and relatively 1).
high intensity; crackles have an impulse-like Santa Catalina (SC) males: SC courtship
quality. The crackle component is generally begins with rapid foreleg flicks followed by
the first song component produced and is pres- body shakes (rapid side-to-side movements)
ent in all observed populations of H. pugillis. during the approach stage of courtship (Mad-
ELIAS ET AL.—SEISMIC SONGS IN HABRONATTUS PUGILLIS 549
Santa Rita Mountains Santa Catalina Mountains Atascosa Mountains
Time Time Time
—
Figures 1-3. Habronattus pugillis song from the Santa Rita, Santa Catalina and Atascosa mountain
populations. 1. Oscillograms of seismic songs. 2. Detail ofoscillograms (boxes in Fig. 1). 3. Spectrogram
of song in Fig. 2. The notation a-c in Figs. 2 & 3 identifies the three seismic components ofmale songs.
dison & McMahon 2000). Afterbody shaking, Atascosa (AT) males: AT locomotory court-
males approach females with vigorous foreleg ship consists of rapid sidling, in which males
flicking. Seismic signals are produced during move in large arcs alternating in direction
leg flicks preceding body shakes as well as with the first pair of legs held continuously
during late display leg flicks. Seismic signals above the ground (Maddison & McMahon
are not produced during a body shake bout but 2000). After sidling, males approach females
occur immediately after body shaking ends. with vigorous foreleg flicking. Seismic signals
Every foreleg flick is coordinated with a seis- are produced during leg flicking and not dur-
mic signal (Fig. 2). SC male seismic songs are ing the sidling display. Every leg flick is co-
long in duration (1.30 ± 0.20 s; n ~ 15) and ordinated with a seismic signal (Fig. 3). AT
occur in two distinct parts (Fig. 2). The first male seismic songs are long in duration (1.85
part consists of a short (0.13 ± 0.16 s; n = ± 0.30 s; n = 15) and occur in three distinct
15), high intensity, broad band (range: 0-3937 parts (Fig. 3). The first part consists ofa short
Hz; peak frequency: 844 ± 349 Hz; /i = 15) (0.11 ± 0.03 s; n = 15), high intensity, broad
“crackle” (labelled “a” in Fig. 2). The sec- band (range: 0-2640 Hz; peak frequency:
ond part consists ofa prolonged long duration 1069 ± 339 Hz; n = 15) “crackle” (labelled
(1.17 ± 0.19 s; n = 15) broadband “rasp” “a” in Fig. 3). The second part consists of a
(range 1: 7-1211 Hz; peak 1 frequency: 185 prolonged (1.07 ± 0.34 s; n = 15) broadband
±198 Hz; range 2: 900-2500 Hz; peak 2 fre- “rasp” (range 1: 0-650 Hz; peak 1 frequency:
quency: 1373 ± 317 Hz; n = \5) (labelled 203 ± 218 Hz; range 2: 530-2010 Hz; peak
“b” in Fig. 2). 2 frequency: 1170 ± 210 Hz; n = 15) (“b”
550 THE JOURNAL OF ARACHNOLOGY
in Fig. 3). The third part consists ofa variable of the forelegs. This character is unique
number of “drones” (3-9; n = 5) of short among all the populations studied (data not
duration (0.04 ± 0.01 s; n = 18) broad band shown).
(range: 0-1850 Hz; peak frequency 371 ±
DISCUSSION
457 Hz; n = 18) signals (“c” in Fig. 3). Broad
band drones also occur along with rasps in the H. pugillis is undergoing diversification
second courtship stage (“b” in Fig. 3) but are driven by sexual selection (Maddison & Mc-
lower in intensity than rasps. Mahon 2000; Masta 2000; Masta & Maddison
Other H. pugillis males: We observed the 2002; Hebets & Maddison 2005; Elias et al.
seismic songs of four additional populations, 2006a). Evidence suggests that sexual selec-
Galiuro (GA), Huachuca (HU), Patagonia tion acting on male secondary sexual charac-
(PA), and Mule (MU) mountains (Fig. 4). De- teristics has driven extensive morphological
tailed measurements were not available for and behavioral divergence between popula-
these populations and we were only able to tions on the sky islands of south eastern Ari-
record songs using the piezoelectric device zona (Maddison & McMahon 2000; Masta &
(see above). It is possible that we were not Maddison 2002). Here we show that the di-
able to observe all song components using this versity observed previously was only a partial
method of sound recording therefore, future picture and is further manifested in the evo-
recordings will be conducted using LDV. lution ofdistinct and stereotyped songs among
Galiuro (GA) visual courtship consists of a different populations.
“first leg wavy circle” where the first legs are H. pugillis songs consist of similar com-
held forward and the tips moved in circles si- ponents, although some populations have
multaneously (but out of phase) (Maddison & more complex songs than others. Males from
McMahon 2000). Periodically the first legs the Santa Rita Mountains have simple songs,
come into phase (sometimes punctuated with consisting ofa single component. Males ofthe
a rapid leg flick). Seismic songs are produced other populations have more complex songs
coincident with the first legs coming into with males from the Santa Catalinas having
phase (with and without leg flicks). Seismic songs consisting of two components, and
songs in the GA population are made of males from the Atascosas having songs con-
crackles (a) and “slow” crackles (s-a) (Fig. sisting of three components. In addition to
4). Slow crackles appear to consist of a series these broad scale differences between popu-
of crackles. Slow crackles have an impulse- lations, temporal and spectral components are
like punctuated quality like crackles and are different between each population. There also
different from rasps as rasps are produced as appears to be variation in the coordination of
a continuous signal. visual and seismic components of courtship.
Huachuca (HU), Mule (MU), and Patagonia Signal evolution involving seismic signals is
(PA) male courtship songs are similar to SC thus potentially occurring along three axes:
male courtship with the notable absence of (1) frequency and temporal characteristics, (2)
body shakes. HU, MU, and PA males ap- song complexity, as measured by the number
proach females with flicking of the first pair of seismic components and, (3) multimodal
oflegs. Seismic songs are produced during leg coordination. Diversification in H. pugillis has
flicks. HU, MU, and PA male seismic songs probably occurred on a small temporal and
occur in two distinct parts, crackles (“a,” Fig. spatial scale suggesting that the song differ-
4) and rasps (“b,” Fig. 4). In the early stages ences between populations are likely due to
of courtship, HU males also add a unique selection and not random effects (Maddison &
component to their display. HU males ap- McMahon 2000; Masta 2000; Masta & Mad-
proach females slowly with the forelegs held dison 2002). Below we discuss some of the
above the ground the entire time. Periodically hypotheses that may drive the observed di-
males open and close their chelicerae during versity of songs.
this approach. Seismic signals are produced Spectral and temporal properties in animal
intermittently as the male slowly approaches songs often relay information about mate
the female. Seismic signals during this portion quality and/or species identity, resulting in
of the display consist of crackle components substantial selection on song properties (An-
and are not coordinated with any movement dersson 1994; Bradbury & Vehrencamp 1998;
ELIAS ET AL.=^SEISMIC SONGS IN HABRONATTUS PUGILLIS 551
Time(seconds)
Figure 4.—Seismic song diversity in the sky islands of southern Arizona. Map of southern Arizona
mountain ranges with outlines representing the lower limit of oak woodland habitat (Brown & Lowe
1982), corresponding to an elevation of —1300-1500 m. Dots show collecting localiti—es for H. pugillis.
Representative songs are shownforpopulationsrecordedusinglaservibrometry (allcaps Atascosa,Santa
Rita, Santa Catalina Mts.) and a piezo-electric sensor (Galiuro, Mule, Patagonia, and Huachuca Mts.).
a-c denote the three possible seismic components of male songs. *S“a denotes a “slow crackle”.
Kotiaho et al. 1998; Parri et al. 2002; Gerhardt tial selective pressures on signal evolution and
& Huber 2002). This may be the case in Ha- on the spectral and temporal characteristics of
&
bronattus songs as well (Elias et al. 2005). signals (Michelsee 1978; Larsen Michelsee
The characteristics of the signaling environ- 1983; Romer 1998; Magal et al. 2000; Elias
ment (e.g., leaves, sand) can also add substan- et al. 2004; Cokl et al. 2005). For example.
552 THE JOURNAL OF ARACHNOLOGY
Elias et al. (2004) demonstrated that seismic formation about multiple aspects of quality.
songs in H. dossenus Griswold 1987 could The observed correlation between population
propagate well on only a subset of available density and signal complexity however is in
substrates, resulting in differential mating suc- the exact opposite direction, as our collection
cess for males across substrates. Elias et al. sites with the most spiders (SR) had the sim-
(2004) went on to suggest that evolution may plest seismic songs (Elias, Hebets & Maddi-
lead to substrate specialization and a tuning of son, unpubl. data). Clearly, future studies fo-
spectral and temporal signal characteristics to cused specifically on testing these hypotheses
the particular signaling substrates available. are necessary.
Similar selective pressures may have lead to A role of antagonistic coevolution has been
differences in the temporal and spectral prop- suggested in the evolution of complex, diver-
erties of songs between different sky island gent courtship displays of H. pugillis (Hebets
populations. & Maddison 2005; Elias et al. 2006a). Under
&
There also appears to be variation in song antagonistic coevolution models (Holland
complexity between sky island populations. Rice 1998), females are expected to evolve
Differences in song complexity among popu- resistance to exploitative male signals thus
lations may be due to differences in the sig- forcing males to elaborate signals that are be-
naling environment between sky islands. Ef- yond the current realm of the female’s resis-
ficacy-based hypotheses of complex signal tance. Following from this, females are pre-
function such as the multiple sensory environ- dicted to prefer males with novel exploitative
&
ments hypotheses (Candolin 2003; Hebets traits over males with local traits for which
Papaj 2005) suggest that multiple signals they have evolved resistance. Under this sce-
evolve when there is variation in the signaling nario, if differences in song complexity in H.
environment so that under some conditions pugillis are being driven by antagonistic co-
some signal components can be transmitted evolution, then we would predict the follow-
effectively when other signal components are ing: SR females should prefer AT and SC
not. Under an efficacy backup hypothesis (He- songs over their own male songs (SR); SC
bets & Papaj 2005), one would predict the AT females should prefer AT songs over SR
signaling environment to be the most variable, songs and their own male songs (SC); and AT
followed by the SC and the SR signaling en- females should not show any preference. He-
vironments. While this possibility remains to bets & Maddison (2005) have already dem-
be explicitly tested, there are no obvious dif- onstrated that SR females prefer AT males,
ferences between sky island signaling envi- and that AT females did not show any pref-
ronments (Masta & Maddison 2002). erences between SR and AT males. In addi-
Differences in song complexity between tion, Elias et al. (2006a) showed that SR fe-
populations may also be due to selection for males preferred AT males only if they could
signal content, such as the need to convey produce seismic signals. Results thus far are
multiple differential messages (Moller & consistent with the hypothesis that differences
Pomiankowski 1993; Johnstone 1996; for re- in song complexity are being driven by antag-
view of content-based hypotheses see Hebets onistic co-evolution.
& Papaj 2005). H. pugillis females mate only Ifwe include songs for which we only have
once (Hebets, unpublished observations), preliminary data, there also appears to be var-
therefore informative signals may be at a pre- iation between the coordination of visual and
mium. Due to differential natural selection seismic signals. Some populations show no
pressure across mountaintop populations, it is multimodal coordination in certain song com-
possible that males from different populations ponents (HU population), while others show
MU
might need to convey different aspects of high degrees of coordination (AT, SC,
quality to local females, resulting in divergent populations). Coordinated signaling in multi-
complex displays. Variable population densi- ple modalities can present animals with mul-
ties could also influence complex signaling tiple advantages including reduced signaling
evolution as increased population density costs (sender), reduced processing costs (re-
could increase male competition for mates, ceiver), and increased information content
potentially resulting in an increase in display (Honey & Hall 1989; Partan & Marler 1999,
complexity as males are forced to provide in- 2005; Rowe 1999; Candolin 2003; Uetz &
ELIAS ET AL.—SEISMIC SONGS IN HABRONATTUS PUGILLIS 553
Roberts 2002; Hebets & Papaj 2005). Differ- mens (Araneae: Salticidae). Animal Behaviour
ences in the importance of coordination and/ 40:884-890.
or differences in the cross-modal interactions Clark, D. L. & G. W. Uetz, 1992. Morph-indepen-
between visual and seismic signal could also dent mate selection in a dimorphicjumping spi-
lead to the differences observed between the der: demonstration of movement bias in female
choiceusing video-controlledcourtshipbehavior.
different populations.
Animal Behaviour 43:247-254.
Although, we have only described a small Clark, D.L. & G.W. Uetz. 1993. Signal efficacy and
proportion of H. pugillis songs, our results the evolution ofmale dimorphismin thejumping
shov/ an interesting parallel with regional spider,Maevia inclemens. Proceedings oftheNa-
song differences in birds (Krebs & Kroodsma tional Academy of Sciences USA 90:11954-
1980). Examining other sky islands in the US 11957.
and Northern Mexico will likely reveal an Cokl, A., J. Presern, M. Virant-Doberlet, G.J. Bag-
even greater diversity of songs and song well &J.G. Millar. 2004. Vibratory signalsofthe
types. Given the extraordinary diversity of harlequin bug and their transmission through
songs observed in this and other studies (Jack- plants. Physiological Entomology 29:372-380.
son 1977; Edwards 1981; Gwynne & Dadour Cokl, A., M. Zorovic, A. Zunic & M. Virant-Dob-
1985; Maddison & Stratton 1988a, 1988b; erlet. 2005. Tuning ofhost plants with vibratory
Noordam 2002; Elias et al. 2003, 2005), we songs ofNezara viridula L (Heteroptera: Penta-
tomidae). Journal of Experimental Biology 208:
propose that jumping spiders are a good sys-
1481-1488.
tem to study the function and evolution of Crane, J. 1949. Comparative biology ofsalticid spi-
songs. ders at Rancho Grande, Venezuela. Part IV: an
ACKNOWLEDGEMENTS analysis of display. Zoologica 34:159-214.
We would like to thank M.C.B, Andrade, Cutler, B. 1988. Courtship BehaviorinHabronattus
captiosus (Araneae, Salticidae). Great Lakes En-
N. Lee, M. Kasumovic, K. Gawera, and the
tomologist 21:129-131.
Hoy, Mason, Andrade, and Hebets Lab
DeVoe, R.D, 1975. Ultraviolet and greenrecepetors
Groups. Funding was provided by NSERC to in principle eyes ofjumping spiders. Journal of
ACM (238882 241419), NIH to RRH General Physiology 66:193-207.
(NlDCROl DC00103), and an NSF IRFP Eakin, R.M. & J.L. Brandenburger. 1971. Fine
(0502239) to DOE. structure of the eyes ofjumping spiders. Journal
LITERATURE CITED of Ultrastructure Research 37:618-663.
Edwards, G.B. 1981. Sound production by courting
Andersson, M. 1994. Sexual Selection. Princeton males of Phidippus mystaceus (Araneae: Salti-
BleUsnt,ivAe.rDs.i,tyRP.rCe.ssH,arPdriinec,etP.onM,cNInetwyrJeer&seyD..S5.99Wiplp-. Edwcaidrades),. PGs.By.ch&e 8R8.:R1.99J-a2c1k4s.on. 1994. The role of
liams. 1981. The spectral sensitivities of identi- experience in the development of predatory be-
fied receptors and the function of retinal tiering haviour in Phidippus regius, a jumping spider
in the principal eyes ofajumping spider. Journal (Araneae, Salticidae) from. Florida. New Zealand
of Comparative Physiology 145:227--239.
Bradbury, J.W. & S.L. Vehrencamp. 1998. Princi- EliJaso,urDn.aOl.,ofEZ.oAo.loHgeybe2t1s:2&69R-.2R7.7.Hoy. 2006a. Fe-
ples of Animal Communication. Sinauer Asso-
male preference for complex/novel signals in a
ciates, Sunderland, Massachusetts. 882 pp.
Broniwtni,esDo.fEt.he&SoCu.tHh.weLsotw.eU..S1.98F2o.resBtioSteircvCicoemGmeun-- Eliassp,ideDr..O.B,ehEa.vAi.orHaelbeEtcso,lRo.gRy.1H7o:y76&5-7A7.1C.. Mason.
eral Technical Report RM-78. Rocky Mountain 2005. Seismic signals arecrucial formalemating
Forest and Range Experimental Station, Fort success in a visual specialistjumping spider (Ar-
Collins, Colorado. aneae: Salticidae). Animal Behaviour 69:931-
Caedolin, U. 2003. The use of multiple cues in 938.
mate choice. Biological Reviews 78:575-595. Elias, D.O., B.R. Land, A.C. Mason & R.R. Hoy.
Clark, D.L. & C.L. Morjan. 2001. Attractingfemale 2006b. Measuring and quantifying dynamic vi-
attention: the evolution of dimorphic courtship sual signals injumping spiders. Journal ofCom-
displays in thejumping spiderMaevia inclemens parative Physiology A: Neuroethology, Sensory,
(Araneae: Salticidae). Proceedings of the Royal Neural, and Behavioral Physiology 192:785-
Society of London Series B-Biological Sciences 797.
268:2461-2465. Elias, D.O., N. Lee, E.A. Hebets & A.C. Mason.
Clark, D.L. & G.W. Uetz. 1990. Video image rec- 2006c. Seismic signal production in a wolf spi-
ognition by the jumping spider, Maevia incle- der: parallel versus serial multi-component sig-
554 THE JOURNAL OF ARACHNOLOGY
nals. Journal of Experimental Biology 209: function: developing a frameworkoftestable hy-
1074-1084. potheses. Behavioral Ecology and Sociobiology
Elias, D.O., A.C. Mason & R.R. Hoy. 2004. The 57:197-214.
effect of substrate on the efficacy of seismic Hill, D.E. 1979. Orientation by jumping spiders of
courtship signal transmission in thejumping spi- the genus Phiddipus (Araneae: Salticidae) during
der Habronattus dossenus (Araneae: Salticidae). the pursuit of prey. Behavioral Ecology and So-
Journal of Experimental Biology 207:4105- ciobiology 5:301-322.
4110. Hoefler, C.D. & E.M. Jakob. 2006. Jumping spiders
Elias, D.O., A.C. Mason, W.R Maddison & R.R. in space: movement patterns, nest site fidelity
Hoy. 2003. Seismic signals in a courting male and the use of beacons. Animal Behaviour 71:
jumping spider (Araneae: Salticidae). Journal of 109-116.
Experimental Biology 206:4029-4039. Holland, B. & W.R. Rice. 1998. Perspective: Chase-
Forster, L. 1982a. Vision and prey-catching strate- away sexual selection: antagonistic seduction
gies in jumping spiders. American Scientist 70: versus resistance. Evolution 52:1-7.
165-175. Honey, R.C. & G. Hall. 1989. Attenuation oflatent
Forster, L. 1982b. Visual communication in jump- inhibition after compound pre-exposure: associa-
ing spiders (Salticidae). Pp. 161-212. In Spider tive and perceptual explanations. Quarterly Jour-
Communication: Mechanisms and Ecological
& nal of Experimental Psychology 41B:335-368.
Significance (P.N, Witt J.S. Rovner, eds.). Jackson, R.R. 1977. Courtship versatility in the
Princeton University Press, Princeton, New Jer- jumping spider, Phidippus johnsoni (Araneae:
Gersheayr.dt, H.C. & E Huber. 2002. Acoustic Com- JacSkaslotni,cidRa.eR).. A1n9i82m.alThBeehbaevhiaovuiror25o:f95c3o-m9m5u7n.icat-
munication in Insects and Anurans: Common ing injumping spiders (Salticidae). Pp. 213-247.
Problems and Diverse Solutions. University of In SpiderCommunication: MechanismsandEco-
GriCshwiocladg,oC.PEr.ess1,98C7h.icAagroe,visIliloinnooifs.th5e31jupmpp.ing spi- eldosg.i)c.alPrSiingcneitfoincaUnnciever(sPi.Nt.yPWrietsts,&PriJn.cSe.toRno,vNneerw,
der genus Habronattus EO. P.-Cambridge (Ara-
Jersey.
neae: Salticidae) with phenetic andcladisticanal- Jackson, R.R., X.J. Nelson & G.O. Sune. 2005. A
yses. University of California Publications in
spider that feeds indirectly on vertebrate blood
Entomology 107:1-344.
Gwynne, D.T & I.R. Dadour. 1985. A new mech- bceyedcihnogossionfgthfeemNaalteionmaolsqAuictaodeesmyasofpreSyc.ienPcreo-s
anism of sound production by courting male USA 102:15155-15160.
jumping spiders (Araneae: Salticidae, Saitis mi- &
Jackson, R.R. S.D. Pollard. 1996. Predatory be-
chaelseni Simon). Zoological Society ofLondon
haviorofjumping spiders. AnnualReviewofEn-
207:35-42.
Harland, D.P. & R.R. Jackson. 2000. Cues by which tomology 41:287-308.
Portia fimbriata, an araneophagic jumping spi- Kotiaho, J.S., R.&V. Alatalo, J. Mappes, M.G. Niel-
der, distinguishesjumping-spiderprey fromother sen, S. Parri A. Rivero. 1998. Energetic costs
prey. Journal of Experimental Biology 203: of size and sexual signalling in a wolf spider.
3485-3494. Proceedings of the Royal Society of London B:
& Biological Sciences 265:2203-2209.
Harclaatnido,n Db.yP.PortiRa.Rf.imJbarcikastao,n.a2s0a0l1t.iciPdresypicdlearsstihfait- Krebs, J.R. & D.E. Kroodsma. 1980. Repertoires
specializes at preying on other salticids: species and geographical variation in bird song. Advanc-
that elicit cryptic stalking. Journal of Zoology es in the Study of Behavior 11:143-177.
255:445-460. Land, M.F. 1969a. Movements of retinae ofjump-
Harland, D.P. & R.R. Jackson. 2002. Influence of ing spiders (Salticidae: Dendryphantinae) in re-
cues fromthe anteriormedial eyes ofvirtual prey sponse to visual stimuli. Journal ofExperimental
on Portia fimbriata, an araneophagic jumping Biology 51:471-493.
spider. Journal of Experimental Biology 205: Land, M.F. 1969b. Structure of retinae of principal
1861-1868. eyes of jumping spiders (Salticidae: Dendry-
Harland, D.P, R.R. Jackson & A.M. Macnab. 1999. phantinae) in relation to visual optics. Journal of
Distances at which jumping spiders (Araneae: Experimental Biology 51:443-470.
Salticidae) distinguish between prey and conspe- Land, M.F. 1985. The morphology and optics of
cific rivals. Journal of Zoology 247:357-364. spider eyes. Pp. 53-78. In Neurobiology of
Hebets, E.A. & W.R Maddison. 2005. Xenophilic Arachnids (EG. Barth, ed.). Springer-Verlag,
mating preferences among populations of the New York.
jumping spider Habronattus pugillis Griswold. Land, M.F. & D.-E. Nilsson. 2002. Animal Eyes.
Behavioral Ecology 16:981-988. Oxford University Press, New York. 221 pp.
Hebets, E.A. & D.R. Papaj. 2005. Complex signal Li, D. & M.L.M. Lim. 2005. Ultraviolet cues affect