b a modiﬁed mole cricket lure and description of

BEHAVIOR

A Modified Mole Cricket Lure and Description of Scapteriscus borellii(Orthoptera: Gryllotalpidae) Range Expansion and Calling Song

in California

ADLER R. DILLMAN,1,2 CHRISTOPHER J. CRONIN,1 JOSEPH TANG,3,4 DAVID A. GRAY,5

AND PAUL W. STERNBERG1,6

Environ. Entomol. 43(1): 146Ð156 (2014); DOI: http://dx.doi.org/10.1603/EN13152

ABSTRACT Invasive mole cricket species in the genus Scapteriscus have become signiÞcant agri-cultural pests and are continuing to expand their range in North America. Though largely subterra-nean, adults of some species, such as Scapteriscus borelliiGiglio-Tos 1894, are capableof longdispersiveßights and phonotaxis to male calling songs to Þnd suitable habitats and mates. Mole crickets in thegenus Scapteriscus are known to be attracted to and can be caught by audio lure traps that broadcastsynthesized or recorded calling songs. We report improvements in the design and production ofelectronic controllers for the automation of semipermanent mole cricket trap lures as well as highlyportable audio trap collection designs. Using these improved audio lure traps, we collected the Þrstreported individuals of thepestmole cricket S. borellii inCalifornia.Wedescribe several characteristicfeatures of the calling song of the California population including that the pulse rate is a function ofsoil temperature, similar to Florida populations of S. borellii. Further, we show that other calling songcharacteristics (carrier frequency, intensity, and pulse rate) are signiÞcantly different between thepopulations.

KEY WORDS Scapteriscus borellii, mole cricket, cricket trap, calling song

Crickets in the family Gryllotalpidae are commonlyknown as mole crickets, aptly named for their largelysubterranean lifestyle and accompanying morphology(e.g., large dactyls for digging), giving them somesimilarities with moles (Fig. 1). Native mole cricketspecies in the United States are typically innocuousand some are rare, presumably owing to the balancebetween population growth, food availability, preda-tors, and disease, among other ecological factors(Vaughn et al. 1993, Frank 2009). Invasive mole crick-etshavebecomesigniÞcant agriculturalpests andhavebeen extensively studied in the southeastern UnitedStates (Walker and Nickle 1981). They are known tocause damage to golf courses, lawns, pastures, and therising turfgrass industry (Liskey 2000). A 1986 esti-mate suggested mole cricket damage and control inAlabama, Florida, Georgia, and South Carolina costUS$80 million combined, annually (Frank and Park-man 1999). The invasive species in the United States,

Scapteriscus abbreviatus Scudder, 1869, Scapteriscusborellii Giglio-Tos, 1894 (previously known as S. acl-etus), and Scapteriscus vicinus Scudder, are thought tohave spread to North America via commercial sailingvessels from South America between 1895 and 1925(Walker and Nickle 1981). S. borellii, the southernmole cricket, is mostly subterranean though it haslong-winged adults that can ßy. It is the most wide-spreadof the invasive species in theUnitedStateswitha known distribution spanning from North Carolina toArizona(Nickle andFrank1988,Walker1984)andhasalso been reported in northern Mexico (Garcõa 2006).Because S. borellii had been detected in Yuma, AZ, itwas assumed that expansion to California was inevi-table (Frank 1994), but conÞrmation of this assump-tion has remained absent from peer-reviewed litera-ture. A recent report from the California Departmentof Food and Agriculture suggested that S. borelliihad expanded its distribution to Downey, CA(http://www.cdfa.ca.gov/plant/PPD/pest_sheets.html-orthoptera), while a recent agricultural brief (July2012) indicates invasive mole cricket presence in Im-perial County, CA (http://ceimperial.ucanr.edu/news_359/Ag_Briefs/).

Aside from their status as pests, biological interest inmole crickets stems from the ampliÞer-like chambersthe males construct to sing their calling song to attractmates and their phonotactic ability. Scapteriscus molecrickets were the Þrst insects found to use phonotaxis

1 Howard Hughes Medical Institute, Division of Biology, CaliforniaInstitute of Technology, 1200 East California Blvd, Pasadena, CA91125.

2 Current address: Department of Microbiology and Immunology,Stanford University, 300 Pasteur Dr, Stanford, CA 94305.

3 Arcadia High School, 180 Campus Dr, Arcadia, CA 91007.4 Current address: Department of Biology, Boston University, 5

Cummington Mall, Boston, MA 02215.5 Department of Biology, California State University, 18111 Nord-

hoff St, Northridge, CA 91330.6 Correspondence to: [email protected].

0046-225X/14/0146Ð0156$04.00/0 � 2014 Entomological Society of America

for mate Þnding where orientation actually occurredduringßight (Ulagaraj andWalker 1973). The songs ofdescribed Scapteriscus mole cricket species are re-markable in a number of ways: they sing in a contin-uous trill, the song is unusually pure, the carrier fre-quency is constant, the level of higher harmonics islow, the calling song is exceptionally loud, at least forS. borellii, and they are sung from specially con-structedhorn-shapedburrows,which contribute to allof these qualities (Bennet-Clark 1970, 1987; Ulagaraj1976; Forrest 1983, 1991).

As signiÞcant agricultural pests, the potential use ofsynthetic or recorded mating calls to lure mole crick-ets into trapswas immediately apparent andput touse.Several components are needed in constructing asound lure and cricket trap: 1) a collection device tocatch and hold the crickets, 2) a sound source thatattracts the crickets, and 3) a controller, providingpower and control of the sound electronics (Walker1996). Sound traps for collecting Scapteriscus molecrickets have been extensively used and altered sincethey were originally implemented in 1973. The Þrsttraps used were relatively simple, composed of a taperecorder, an ampliÞer, and a speaker, suspended overa large funnel attached to a jar (Ulagaraj and Walker1973). The tape recorders were used to broadcastÞeld-recorded songs or synthetic calling songs madeand tape-recorded in the lab. Nearly a decade afterthese initial traps, improvements were made for boththe collection device and the lure, which were calledOldacre traps (after their designer, William Oldacre[Walker 1982]). The improvements allowed for lim-ited automation and lead to less supervised, less cum-bersome, and more efÞcient cricket traps (Walker1982). Several years later, themole cricket sound traps

were further improved, as the design of the lure wastailored for increased economy and versatility, in-creased automation, and made amenable to larger-scale production through commercial contracting.Thesealterations led to the “artiÞcial cricket”or “Mansunit” (named for the designer, Bernard Mans), whichis still used as the lure for traps in recent studies(Walker 1990, 1996; Frank and Walker 2006; Thomp-son et al. 2007). Further improvements to the collec-tion device, making use of large plastic wading pools,have made the sound traps more affordable, quickerand easier to assemble, temporary at each site (whichis appreciated if not demanded by golf course super-intendents), and able to maintain live specimens for7 d, thus requiring even less supervision (Thompsonand Brandenburg 2004).

We explored the possibility of reproducing theMans-design artiÞcial cricket and found the literatureon the collection part of the device, particularly therecent improvement of plastic wading pools, to beclear and easy to follow (Thompson and Brandenburg2004). However, we found the complexity of the as-sembly language software, the seemingly incompletewritten design, and the unavailability of newly pro-duced models of the Mans-design artiÞcial cricketdiscouraging. We determined to redesign the control-ler so that rather than being commercially ordered itcould be constructed in lab with minimal to no engi-neering background. We wrote the underlying sourcecode for the controller in an open-source language,allowing individual researchers to easily modify theoperating conditions, thus making this trap more ac-cessible to the biological community.

Here we describe the design and construction of anew audio cricket trap, which we used to survey theRioHondoGolf course inDowney,CA.We report thepresence of S. borellii in southern California, conÞrm-ing previous assumptions that this range expansionwould occur (Frank 1994), and we describe the Cal-ifornia population of S. borellii and analyze elementsof the male calling song.

Materials and Methods

Construction of a Semipermanent Trap. Three ma-jor components are involved in the construction of anautonomous mole cricket audio trap (Walker 1982):1) a catching device, 2) sound source, and 3) con-troller.

1. Catching device: We constructed a wading poolcatching device, the latest improvement to molecricket audio lure traps, as previously described(Thompson and Brandenburg 2004; Fig. 2). Brießy,two large plastic wading pools (�5 feet or 1.5 m indiameter) are used, one suspended over the otherby four evenly distributed 2- by 6-inch nominalwooden spacers (actual dimensions 1.5 by 5.5inches or 38 by 140 mm), which prevent saggingand distortion of the top pool over time. The poolsand spacers were secured with a bolt, washer, andnut for each of the four spacers. The top pool had

1 cm

left forleg

1 cm

Fig. 1. Pictures of identifying features of S. borellii fromDowney, CA.

February 2014 DILLMAN ET AL.: MOLE CRICKETS IN CALIFORNIA 147

12 evenly spaced 13.5-cm-diameter holes cutthrough its bottom, allowing crickets to fall into thebottom pool, which was Þlled with moist sand.Small drainage holes and slits were drilled or cutinto the bottom of the bottom wading pool to allowdrainage.Weatherproof speakersweremountedona 6-foot (183-cm)-long and 1- by 4-inch nominalwooden fence board (actual dimensions 0.625 by3.375 inch or 16 by 86 mm) along with a weather-proofed photoresistor for automated operation ofthe controller and audio playback. The fence boardwas then mounted across the top of the upperwading pool, being secured to two diametricallyopposed spacers by the bolts used to secure thespacers to the pools (Fig. 2).

2. Sound source: We explored several options for am-plifying and broadcasting the cricket mating call,but for simplicity we elected to use a commerciallybuilt ampliÞer and speaker rig intended for useoutdoors on a motorcycle or snowmobile (modelPLMCA20, Pyle Audio, Brooklyn, NY). The se-lected soundsystemfeaturedanauxiliary input jackcapable of accepting output from an external audiosource. The accompanying speakers were alreadywaterproof, eliminating the need for furtherweather protection such as a plastic bag, whichcould distort or otherwise affect the broadcastedsignal (Fig. 2 and 3A). A Þeld recording of a malecalling song was looped and used for the trap lure(Supp. material [online only]).

3. Controller: The electronic system that controlledthe timing and function of our cricket lure, thecontroller, was constructed to prioritize certainfunctional features such that the trap would beginplaying the recorded mating call at dusk, play themating call at the predetermined optimal soundpressure level of 106 dB (measured at 15 cm fromthe audio source; Walker and Forrest 1989), stopplaying after 120min, reset itself to repeat this cycleon the following day, and be capable of performingthese functions unsupervised for at least a week ata time.

We selected anArduinoUnomicrocontroller boardto run the various functions of our audio lure (http://arduino.cc/en/). In addition to ease of use, the open-source nature of ArduinoÕs hardware and softwaremade it ideal for our purposes of developing a systemeasily accessible to the biology research community.We attached a Wave Shield (model v1.1, AdafruitIndustries, NY; http://www.ladyada.net/make/waveshield/) to the Arduino microcontroller to han-dle playback of audio Þles stored on a secure digital(SD) card (http://www.adafruit.com/products/94;Fig. 3A). An external photoresistor and indicator LED(bothwaterproofed)were interfaced to thecontrollerthrough separate circuit boards (Fig. 3A). There wasno electrical outlet at our sampling site, which re-quired our design to be battery powered. A 35 ampÐhour (Ah), 12 V battery (model PS12350 NB, Power-Sonic, San Diego, CA) was used to power thecontroller and sound source for the semipermanentpool trap. This battery was recharged periodicallyusing a ßoating charger (model FC-12250, 12 VDC1.5A output, PHC Enterprise, Torrance, CA), but tomaximize available battery power, we built a simplerelay-based switching circuit that the Arduino acti-vated to power the ampliÞer only when playing audio.Nonweatherproofed components like the Arduino,Wave Shield, and battery were housed in a coveredplastic container setnext to thewadingpools (Fig. 3A)for protection from the elements while weather-proofed elements like the speakers, LED, and photo-resistor were left exposed at the trap. All electroniccomponents for the controller assembly were pur-chased from local retailers and from online merchants(Orvac Electronics, Pasadena, CA; FryÕs Electronics,Los Angeles, CA; and Amazon.com).

We programmed the controller through the Ar-duino Integrated Development Environment usingthe open source Arduino programming language, ahigh-level language similar to the C programming lan-guage (http://arduino.cc/en/main/software). In pro-gramming the controller, we wrote the code with thegoal that all functional features could be easily ad-

Fig. 2. The design of the semipermanent wading pool trap.

148 ENVIRONMENTAL ENTOMOLOGY Vol. 43, no. 1

justed to suit different needs (Supp. material [onlineonly]).

Aschematicdrawingoftheelectronics isshowninFig.3B. For additional methods, controller details, and thefull programming code see Supp.material (online only).

Construction ofHighly PortableTraps.Thedesignsof the portable traps were much simpler than that ofthe semipermanent trap, with ease of portability by asolo researcher of particular importance. Because the

portable traps were intended to be delivered, set up,run, brokendown, and returned to in-laboratory storageeach trappingevening, anautonomouscontrollerwasnolonger necessary. We developed Þve portable trap de-signs thatwe refer to as the “Benjamin,” “Cross,” “Sheet,”“Table,” and “Umbrella” designs (Fig. 4).

1. Catching device: These designs vary from a sophis-ticated inverted umbrella-like design requiring

A Controller

B Controller cirucuit schematic

Controller

Photosensor interface

Relay circuit andLED interface

RemotePhotoresistor

10kΩ

PotentiometerTrimming

Resistor1k Ω

RemoteLED

Power toAmplifier ( + )

Power toAmplifier ( - )

Transistor2N2222

1N4004Diode

Relay

Relay Board

220ΩResistor

12V Battery

N/CIO REF

SCLSDA

RSTAREF

A0A1A2A3A4A5

Analog Input

D11

D10D9D8

D7D6

D5D4

D3 PWM

TXRX

PWMPWM

PWMPWM

PWM

D2

D1

D13D12

D0

Dig

ital I

nput

/Out

put

Power

withWave Shield

Arduino

GND

Use

d by

Wav

e S

hiel

dU

sed

by W

ave

Shi

eld

LightSensitivityAjdustmentBoard

3V3 5V Vin

Fig. 3. (A) The controller and electronics of the wading pool trap, housed in a plastic weatherproof container. (B)Controller circuit schematic depicting Arduino microcontroller with Wave Shield, separate Light Sensitivity Adjustment andRelay boards, and remote photoresistor, indicator LED, and audio ampliÞer connection.


well-calculated measurements during fabricationtoabed sheetwithaholecutout, stapled toaplasticcontainer. Full details of the designs are available inthe Supp. material (online only).

2. Sound source: Two different sound sources wereused interchangeably to operate the Þve trap de-

signs, both using the same 2-h audio recording asthe semipermanent trap.TheÞrst designplayed the2-h audio recording from a small and affordableMP3 player (model SMP2012 2GB, Sylvania, Dan-vers, MA) connected to an ampliÞer and speakersystem identical to that used for the pool trap

A The “Benjamin” portable trap

B The “Cross” portable trap

C The “Sheet” portable trap

D The “Table” portable trap

E The “Umbrella” portable trap

F MP3 portable controller G SD card portable controller

Fig. 4. The portable trap designs. (A) “Benjamin.” (B) “Cross.” (C) “Sheet.” (D) “Table.” (E) “Umbrella.” (F) MP3portable controller. (G) SD card portable controller.


(model PLMCA20, Pyle Audio) but that used onlyone of the two speakers. The other audio designused an ampliÞer with a built-in SD card reader(model PLMCA25U, Pyle Audio) to play the audiorecording stored on an SD card through a singlespeaker fromaPyle PLMCA20 system. These audiosystems were each powered by a 5 Ah, 12 V re-chargeable, sealed lead acid battery (model PS-1250 F1, Power-Sonic, San Diego, CA), which weregularly rechargedusing aßoating charger (modelFC-12250, 12 VDC 1.5A output, PHC Enterprise,Torrance, CA). These portable traps were set upand the lures manually turned on for 1.5Ð2 h, tocoincide with live mole cricket calling songs at theRio Hondo Golf Course. Lures were manually shutoff after the resident mole crickets stopped callingfor the night. Any crickets found in or around thetraps were then counted and collected and thetraps were taken down and stored until the nexttrapping evening.

Traps were used at Þve different locations (Fig. 5),chosen owing to high visible cricket activity and theadvice of the golf course superintendent who hadbeen dealing with mole cricket damage since 2006.Portable trapswere set up and used on seven differentnights from June to July of 2011. Each of the fourportable trap locationswas used at least four times andas many as seven (Fig. 6; Table 1 and Supp. Table 1[online only]).

Mole Cricket Song Recording and Analysis. South-ern mole crickets, S. borellii, were collected from theRio Hondo Golf Club, Downey, CA, from April to Julyof 2011. Collecting was done using a semipermanentand highly portable audio lure traps. The semiperma-

nent trapwasprogrammed toplay an audio lure for 2hnightly from May through July of 2011. It was set upinside the golf course nursery to prevent possible theftor human disturbance of the trap and because thenursery was one of the highest activity areas noticedby the superintendent. We checked the trap period-ically (semipermanent trapwas checkedevery 4Ð12d;Table 2) to see how many crickets it had trapped, torecharge thebattery, and to verify that the automationwas working properly. Captured crickets were used inhost-range and -preference assays using entomopatho-genic nematodes. These experiments were publishedpreviously (Dillman et al. 2012).

We audio recorded numerous individual male molecricket calling and courtship songs in the Þeld at theRio Hondo Golf Course. Songs were recorded usingthe onboard microphone in an Apple MacBook Pro(model MC374*/A, Apple, Cupertino, CA) with theAudacity software package (http://audacity.source-forge.net/) with AudacityÕs default recording settings(Stereo, 44100Hz, 32-bit ßoat). Individual males wererecorded for 1.5Ð2 min each. Recordings were takenbypositioning themicrophone�15cmfromtheopen-ing of a singing maleÕs burrow. Sound pressure mea-surements of calling songs were taken at 15 cm fromthe burrow opening using a sound-level meter (model33-2055, RadioShack, Fort Worth, TX) with the me-terÕs “C-weighting” and “Fast response time” settings.Calling songs were analyzed using the software pack-age Raven Pro 1.4 (Cornell Laboratory of Ornithol-ogy, Ithaca, NY). We measured a deÞned set of tem-poral parameters, which can be divided into thefollowing trill elements for the calling songs: trilllength (TL), duration of the break in the trill (BD),

1

2

3

4

5Fig. 5. Sampling locations used at the Rio Hondo Golf Club in Downey, CA.


the length of the pulses in the trill (TPL), the intervalbetween the pulses in the trill (TPI), and the numberof pulses in the trill element (P/T; Fig. 7; Simmons etal. 2010).

Statistical Analysis. S. borellii calling song metrics(intensity, carrier frequency, and pulse rate) werecompared between the Florida and California popu-lations using an unpaired t-test with WelchÕs correc-tion using the calculated mean, standard deviation,and sample size rather thanusing rawdatabecause theraw data were not available for the Florida population(Ulagaraj 1976). Regression analysis was performedon the raw calling song data from the California pop-ulation. All analyses were carried out using GraphPadPrism version 6.0a for Mac OS X (GraphPad Software,San Diego, CA, http://www.graphpad.com).

Results

Pool Trap. The controller and light sensor assemblyproved capable of appropriately responding to dusk;speciÞcally, it turned the audio lure on and off asdesired for the entire course of our trapping experi-ment (53 d), being checked only periodically forcricket collection and battery charging (Supp. Table2 [online only]). Electrical current drawwas a seriousconcern, especially because the trap needed to func-tion autonomously for at least 7 d to minimize super-vision. The controller drew 61 milliamps (mA) at 12.5

Sampling Location

Cric

kets

cau

ght

Loca

tion 2

Loca

tion 3

Loca

tion 4

Loca

tion 5

-5

0

5

10

15

Date of sampling

Cric

kets

cau

ght

6/13

6/16

6/20

6/21

6/27 7/5 7/1

1-5

0

5

10

15

A

C

B

Benjam

inCros

sShe

et

Table

Umbrella

-5

0

5

10

15

Trap type

Cric

kets

cau

ght

Fig. 6. Trapping data for the Þve mobile trap designs. (A) Trapping data by trap type. (B) Trapping data by samplinglocation. (C) Trapping data by date of sampling.

Table 1. Number and gender of S. borellii caught in and aroundthe different audio traps used

Trap type Females caught Males caught No. days tested

Semi-permanent 16 0 53Highly portable 48 13 15

SpeciÞc details of where crickets were caught and how many dayseach trapwas used at thedifferent locations correlatingwithFig. 5 arerecorded in Supp. Table 1 (online only).

Table 2. Characteristics of male S. borelli calling songs

CharactersCalifornia

mean � SD(N � 37)

Floridamean � SD

(N � 52, 70, 70)

Unpairedt-test

Intensity (db) 74.73 � 2.36 68.5 � 12.8 0.0011Carrier frequency (Hz) 2783.8 � 97.7 2620 � 140 �0.0001Pulse rate (pulses/s) 50.54 � 3.97 54.7 � 4.6 �0.0001Trill length (TL) 16.86 � 16.11 Ð ÐTrill break duration (BD) 0.525 � 0.5 Ð ÐTrill pulse length (TPL) 0.013 � 0.001 Ð ÐTrill pulse interval (TPI) 0.007 � 0.001 Ð ÐPulses per trill (P/T) 859.1 � 884.3 Ð Ð

Florida data from Ulagaraj (1976). Recorded in the Þeld (seeMethods).


VDC, even when the lure was not broadcasting andthis increased to 130 mA when playing audio at fullvolume (full Wave Shield volume). The ampliÞerbroadcasting theWave ShieldÕs output through speak-ers at the previously determined optimal sound pres-sure level of 106 dB (measured 15 cm in front of thespeaker; Walker and Forrest 1989) drew �450 mA.The 35-Ah battery proved capable of maintainingthe full function of the trap and lure for as many as12 consecutive days (Table 2), and potentially lon-ger, though this may not be a good strategy forcollecting S. borellii for any downstream applica-tions, as they are cannibalistic (Thompson andBrandenburg 2004, Thompson et al. 2007). The pooltrap lured and trapped 16 female crickets during its53 d of operation (Table 1 and Supp. Table 1 [onlineonly]).

Portable Traps. The Þve different portable trapsused each caught between 6 and 15 mole crickets infour to Þve nights of use, for a total of 48 female and13 male S. borellii adult crickets (Table 1). All of thesedesigns caught at least six crickets in the four to Þvenights of sampling that each trap was used (Fig. 6;Table 1 and Supp. Table 1 [online only]). Some siteswere more productive than others, with site 3 beingthe most productive, yielding 39 crickets, 64% of ourportable trap total, caught over seven nights of sam-pling with different traps (Fig. 6; Table 1 and Supp.Table 1 [online only]). Site 5 was the least productiveonly leading to a single cricket caught in four nights oftrapping. Some nights were better for catching crick-ets than others, and this could explain the difference

in site productivity observed. More sampling days us-ing the portable traps would be required to makestatistically meaningful conclusions about the effectsof trap type, location, and sampling day (Fig. 6).Despite previous reports of other crickets some-times being caught in these audio lure traps (Ula-garaj and Walker 1973), we caught only S. borelliicrickets in these traps. Many of the crickets wecaught were found near (within 10 feet) the traprather than in it (37 in traps and 24 near; Supp. Table1 [online only]). Three of the 13 males caught wereactually inside a trap, with the rest being caughtnear traps.

S. borellii Song Analysis. As noted in Ulagaraj(1976), the S. borellii calling song is a long bell-shapedtrill with three sections: 1) a few seconds of high-amplitude pulses, 2) followed by a period of pulseswith decreasing amplitude, and 3) continuing withpulses at the decreased amplitude for up to severalminutes until the end of the trill (Fig. 7). Table 2provides the calling song elements we analyzed. Wecompared calling song elements from the Downey,CA, S. borellii population with those reported for an S.borellii population near Gainesville, FL (Ulagaraj1976). The measurements of intensity, carrier fre-quency, and pulse rate of California S. borellii weresigniÞcantly different from those elements of FloridaS. borellii using a corrected unpaired t-test, indicatingthat the California population sampled in this studycalled at slightly higher intensity, higher carrier fre-quency, and slower pulse rate (Table 2). We alsofound, using regression analysis, that the pulse rate of

24

0.0048

1216

-4-8

-12-16kusec.

kHzsec.

26 28 30 32 34 36 38 40 42 44 46 48 50 52 54

101214

2.783 kHz

1618

8642

24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54

A

B

C TL

TPI TPL

BD

Fig. 7. Representative S. borellii calling song analysis from the California population. (A) Oscillogram of the S. borelliicalling song trills. (B) Audiospectrogram corresponding to the above oscillogram. (C) Stylized oscillogram of the S. borelliicalling song, illustrating the parameters measured. The mean durations and values of the various song elements are presentedin Table 2. TL, trill length; BD, trill break duration; TPI, trill pulse interval; TPL, trill pulse length.


the calling song is a function of soil temperature (Fig.8), which agrees with the earlier Þndings of Ulagaraj(Ulagaraj 1976). The slope of our regression (1.465 �0.233 p/s per degree) did not differ signiÞcantly fromthe work of Ulagaraj (1.4 p/s per degree).

We were able to record several courtship songs forthe California S. borellii population and found theseries of short rhythmic chirps to be just as previouslydescribed for the Florida population (Ulagaraj 1976).However, we did observe a different and previouslyunreported alternative state within the courtship call,

characterized by a distinct change in the number ofpulses.Thepreviouslydescribedcourtship song showschirps being progressive and abruptly cutting off. Thissecond state within the courtship song changes, withthe chirps following a different pattern of ascendingand descending in relative amplitude rather than end-ing abruptly at or near peak amplitude as happensduring the rhythmic chirping state of the courtshipsong (Fig. 9).

Discussion

We report the Þrst known establishment of S. borel-lii in California, conÞrming previous fears that thisspecies would continue to spread West (Nickle andFrank 1988, Frank 1994). To study these crickets, wehave redesigned and updated the lure and controllerfor an acoustic mole cricket trap. The previous designwas developed in 1989 and available only throughcommercial production (Walker 1990, Thompson andBrandenburg 2004). Our improved design uses easilyavailable components, open access software, and canbe built “in-house” by tinkering biologists, no engi-neering background required. We have shown thatthis design can operate unsupervised for at least 12 dand is easily modiÞable to suit individual needs. Inaddition to the improvements of the semipermanenttrap design, we have designed and used highly por-table traps of varying complexity from a sophisticatedinverted umbrella design to a simple bed sheet stapledto a plastic box, all of which we have successfully used

20 22 24 26 2835

40

45

50

55

60Pu

lse

rate

(pul

ses/

sec)

Temperature

Y = 1.465*X + 15.65

Fig. 8. The effect of soil temperature on the pulse rateof 37 S. borellii calling songs. Linear regressionwas calculatedon all individual values, with vertical lines showing the rangeof mean values.

1:13.5 1:14 1:14.5 1:15 1:15.5 1:16 1:16.5

1:13.5 1:14 1:14.5 1:15 1:15.5 1:16 1:16.5

0.00

0.00

5

10

15

-2-4

42

kum:s

kHz

m:s

A

B

0.00-2-4

42

kum:s

C

D

1:18.5 1:19 1:19.5 1:20 1:20.5 1:21 1:21.5 1:22 1:22.5 1:23 1:23.5

1:19 1:19.5 1:20 1:20.5 1:21 1:21.5 1:22 1:22.5 1:23 1:23.50.00

5

10

kHz

Fig. 9. Representative S. borellii courtship song from the California population. (A) Oscillogram of courtship song. (B)Audiospectrogram of the corresponding oscillogram. (C) Oscillogram of courtship song illustrating a new phase of the song,with variable pulses per chirp. (D) Audiospectrogram of the corresponding oscillogram.


in the Þeld to catch mole crickets (Table 1 and Supp.Table 1 [online only]).

At present the Downey S. borellii population seemsto be quite small and is subject to rigorous controltreatments by the landscaping crew and superinten-dentof theRioHondoGolfClub.A recent agriculturalbrief suggests that Scapteriscus mole crickets are alsoin Imperial County (http://ceimperial.ucanr.edu/news_359/Ag_Briefs/), though this Þndinghasnot yetbeen reported in peer-reviewed literature. Previousresearchers assumed that the presence of S. borellii inYuma, AZ, ensured their eventual spread to California(Frank 1994), which we now report to be true. Nickleand Frank (1988) suggested that Mexico could sooninherit the pest insect S. borelli from theUnited States,which also seems to have happened (Garcõa 2006).Single S. borellii individuals are known to disperse byßying and have been recaptured �700 m from releaseexperiments (Ulagaraj 1975). The factors that causethese dispersive ßights remain unknown (e.g., over-crowding, availability of food, availability of mates,etc.) The agricultural impact of these mole crickets inCalifornia is minuscule at present, but their potentialto expand into more areas of California and Mexico isforeboding. The seemingly limited distribution of S.borellii in California suggests that an aggressive con-trol program could prevent population growth or fur-ther spread (Nickle and Frank 1988; Frank 1994, 2009;Frank and Walker 2006). However, the presence of S.borellii in Yuma for at least 25 yr and their morerecently observed expansions into certain parts ofCalifornia and Mexico, indicate that this pest willlikely continue to spread if unchecked. The expansionof S. borellii in the West also highlights the possibilityof other ßying mole cricket pests, such as S. vicinus, toexpand their distribution further West as well.

To better characterize the Downey population of S.borellii,wehave reportedhere several elements of themale calling song (Table 2). We have shown that, likeFlorida S. borellii populations, pulse rate of the callingsong is a function of soil temperature (Fig. 8). How-ever, all other calling song elements compared (in-tensity, carrier frequency, and pulse rate) were sig-niÞcantly different between Florida and Californiapopulations(Table2). It is not currently clear tous thesigniÞcance of these calling song differences. Perhapsthese differences reßect the different climates ofsouthern Florida and southern California, or theycould be suggestive of evolutionary changes betweenthe populations. Assuming the Downey population isderived from expanding progeny of one of the knownfounding events in the West (Walker 1984), the Flor-ida population and the Downey population of S. borel-lii have been separated for �80Ð110 yr. The differ-ences in calling song that we have identiÞed comparerecordings made decades apart with different (butbroadly comparable) recording technologies, and thedifference we have identiÞed highlights the need forfurther investigations of calling song variation be-tween populations.

Acknowledgments

We thank John Rodriguez and the Rio Hondo Golf Clubfor the cooperation and help in our sampling efforts; JohnDeModena for fruitful discussions, advice, and critical read-ing of the manuscript; Benjamin Cronin for trap design dis-cussions and recommendations; and J. Howard Frank formany suggestions, discussion, critical reading of the manu-script, and sharing hismole cricket expertise.We thankBrianAnderson,MichaelE.Castillo, andEstherThompson forhelpwithmole cricket collection.Wealso appreciate the timeandsuggestions of anonymous reviewers. This work was sup-portedbyaNational InstitutesofHealth(NIH)UnitedStatesPublic Health Service Training Grant (5T32GM07616) toA.R.D. and by the Howard Hughes Medical Institute (withwhich P.W.S. is an investigator).

References Cited

Bennet-Clark, H. C. 1970. The mechanism and efÞciency ofsound production in mole crickets. J. Exp. Biol. 2: 619Ð652.

Bennet-Clark, H. C. 1987. The tuned singing burrow ofmole crickets. J. Exp. Biol. 128: 383Ð409.

Dillman, A. R., M. L. Guillermin, J. H. Lee, B. Kim, P. W.Sternberg, and E. A. Hallem. 2012. Olfaction shapeshost-parasite interactions in parasitic nematodes. Proc.Natl. Acad. Sci. U.S.A 109: E2324ÐE2333.

Forrest, T. G. 1983. Calling songs and mate choice in molecrickets, pp. 185Ð204. In D. T. Gwynne and G. K. Morris(eds.), Orthopteran Mating Systems: Sexual Competitionin a Diverse Group Of Insects. Westview Press, Boulder,CO.

Forrest, T. G. 1991. Power output and efÞciency of soundproduction by crickets. Behav. Ecol. 2: 327Ð338.

Frank, J. H. 1994. Biological control of pest mole crickets,pp. 343Ð352. InD.Rosen,F.D.Bennett, and J. L.Capinera(eds.), Pest Management in the Subtropics. BiologicalControl-A Florida Perspective. Intercept, Andover, Eng-land, United Kingdom.

Frank, J.H. 2009. Steinernema scapterisci as a biological con-trol agent of Scapteriscus mole crickets, pp. 115Ð131. InA. E. Hajek (ed.), Use of Microbes for Control and Erad-ication of Invasive Arthropods, vol. 6. Springer, Amster-dam, The Netherlands.

Frank, J. H., and J. P. Parkman. 1999. Integrated pest man-agement of pest mole crickets with emphasis on thesoutheastern USA. Integr. Pest Manage. Rev. 4: 39Ð52.

Frank, J. H., and T. J. Walker. 2006. Permanent control ofpestmole crickets (Orthoptera: Gryllotalpidae: Scapteris-cus) in Florida. Am. Entomol. 52: 138Ð144.

Garcıa, E. R. 2006. An annotated checklist of some or-thopteroid insects of Mapimi biosphere reserve (Chihua-huan desert), Mexico. Acta Zool. Mex. 22: 131Ð149.

Liskey, E. 2000. TurfÕs most (un)wanted pests. GroundsMaint. 35: 18.

Nickle, D. A., and W. Frank. 1988. Pest mole crickets, Scap-teriscus acletus (Orthoptera,Gryllotalpidae), reportedes-tablished in Arizona. Fla. Entomol. 71: 90Ð91.

Simmons, L.W.,R.M.Tinghitella, andM.Zuk. 2010. Quan-titative genetic variation in courtship song and its cova-riation with immune function and sperm quality in theÞeld cricketTeleogryllus oceanicus.Behav. Ecol. 21: 1330Ð1336.

Thompson, S. R., and R. L. Brandenburg. 2004. A modiÞedpool design for collecting adult mole crickets (Or-thoptera: Gryllotalpidae). Fla. Entomol. 87: 582Ð583.


Thompson, S. R., R. L. Brandenburg, and G. T. Robertson.2007. Entomopathogenic fungi detection and avoidanceby mole crickets (Orthoptera: Gryllotalpidae). Environ.Entomol. 36: 165Ð172.

Ulagaraj, S. M. 1975. Mole crickets-Ecology, behavior, anddispersal ßight (Orthoptera-Gryllotalpidae-Scapteriscus).Environ. Entomol. 4: 265Ð273.

Ulagaraj, S. M. 1976. Sound production in mole crickets(Orthoptera-Gryllotalpidae-Scapteriscus).Ann.Entomol.Soc. Am. 69: 299Ð306.

Ulagaraj, S. M., and T. J. Walker. 1973. Phonotaxis of crick-ets in ßight: attraction ofmale and female crickets tomalecalling songs. Science 182: 1278Ð1279.

Vaughn, C. C., S. M. Glenn, and I. H. Butler. 1993. Char-acterization of prairie mole cricket chorusing sites inOklahoma. Am. Midland Nat. 130: 364Ð371.

Walker, T. J. 1982. Sound traps for sampling mole cricketßights (Orthoptera:Gryllotalpidae:Scapteriscus).Fla.En-tomol. 65: 105Ð110.

Walker,T. J. 1984. Biologyofpestmolecrickets: systematicsand life cycles. Univ. Fla. Agric. Stn. Bull. 846: 3Ð10.

Walker, T. J. 1990. Annual report on mole cricket research,pp. 6Ð17. In: Annual Report 89Ð90, Department of En-tomology and Nematology. University of Florida, Gains-ville, FL.

Walker, T. J. 1996. Acoustic methods of monitoring insectpests and their natural enemies, pp. 245Ð257. In D.Rosen, J. L. Capinera, and F. D. Bennett (eds.), PestManagement in the Subtropics: Integrated Pest Man-agement-A Florida Perspective. Intercept Ltd., An-dover UK.

Walker, T. J., and T. G. Forrest. 1989. Mole cricket phono-taxis: effects of intensity of synthetic calling song (Or-thoptera: Gryllotalpidae: Scapteriscus acletus). Fla. Ento-mol. 72: 655Ð659.

Walker, T. J., and D. A. Nickle. 1981. Introduction andspread of pest mole crickets: Scapteriscus vicinus andScapteriscus acletus reexamined. Ann. Entomol. Soc. Am.74: 158Ð163.

Received 17 May 2013; accepted 13 December 2013.


b a modiﬁed mole cricket lure and description of

Documents