ABSTRACT

HERTL, PETER THOMAS. The life cycle and development of Scapteriscus vicinus Scudder and S. borellii Giglio-Tos (Orthoptera: Gryllotalpidae) in southeastern North Carolina. (Under the direction of Rick Lynn Brandenburg.)

The development of the tawny (Scapteriscus vicinus Scudder) and southern mole cricket (S. borellii Giglio-Tos) was quantified in southeastern North Carolina from field-collected samples. Nymphs were sampled weekly using the soapy water flush technique during the summers of 1993–1997 at nine golf courses to compile a developmental data base from a total of 20 site-years. Pronotal length was used to assign the nymphs to size classes, and counts were summarized on a m2 basis. The two smallest size classes were equated to the first and second instar and counts were used to quantify the timing of 25, 50 and 75% peak and cumulative abundance and estimate the timing of oviposition and hatch. Flight behavior was studied concurrently at two sites using paired acoustic calling traps to capture flying adults. Timing of flight was compared to timing of nymph abundance. Flight counts were also compared to nymph counts before and after flight at both sites, and with damage levels occurring later in the season at one site. The relationship between soil moisture and damage was studied (1995–1997) in the field at one site and the relationship between soil moisture and oviposition was studied in a series of greenhouse experiments. Soil moisture was found to significantly affect both surface activity and oviposition. On-site soil degree-day accumulations and rainfall data from regional weather stations were examined to determine their relationship to the timing of development and flight. Soil degree-days were correlated with the timing of cumulative nymph abundance, however, date quantified timing better than degree-days. A statistically significant correlation between the timing of development and both soil degree-days and rainfall was found. Differences in annual degree-day accumulations and a soil moisture-related delay in oviposition documented in the greenhouse experiments are believed responsible for observed differences in annual development. Management implications of the research are discussed. THE LIFE CYCLE AND DEVELOPMENT OF SCAPTERISCUS VICINUS SCUDDER AND S. BORELLII GIGLIO-TOS (ORTHOPTERA: GRYLLOTALPIDAE) IN SOUTHEASTERN NORTH CAROLINA

by

PETER THOMAS HERTL

A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the Degree of Doctor of Philosophy

ENTOMOLOGY

Raleigh

2003

APPROVED BY:

______

______

______Chair of Advisory Committee DEDICATION

This work is dedicated to the memory of Dr. Jack E. Bailey who served on my graduate committee as a friend, colleague, mentor and advisor until his untimely death in 2002. Jack always encouraged me in his gentle way and offered guidance and advice on many fronts, both professional and personal. He always found time to see me, even when he had a heavy schedule and had other important and pressing things to do. His experience, counsel, encouragement and support were much missed in the final days of this project. His passing has helped me to realize the short and fragile nature of life, perhaps a more important lesson than any learned in the classroom or field. It is with deep regret that he is not here to share in the joy of this work’s completion. He will be missed, but never forgotten.

ii BIOGRAPHY

Peter Thomas Hertl was born on 11 August 1954 in Allentown, Pennsylvania. His family moved to Towson, Maryland where he completed his secondary education at Towson

Senior High School, graduating in 1973. In September of 1973, he began undergraduate studies in Biology at Randolph-Macon College in Virginia. He transferred to North

Carolina State University (NCSU) in 1975 and completed a Bachelor of Science degree in

Wildlife Biology in 1978. Graduate studies in Acarology leading to a Master of Science degree in Entomology were completed at NCSU in the spring of 1986. Since that time he has been employed as a Technician, and later as a Research Assistant in Peanut and

Turfgrass Entomology in the Department of Entomology at the University. He is an avid caver and has traveled extensively in Mexico and Venezuela. He began doctoral studies in

Entomology at NCSU in the fall of 1993 and is currently a member of the North Carolina

Entomological Society and the Entomological Society of America.

iii ACKNOWLEDGMENTS

I would like to extend my sincere thanks to Dr. Rick L. Brandenburg for his patience, guidance and support of these graduate studies. The additional generous support, advice and critical review of manuscripts by committee members Drs. Mary E. Barbercheck, P. Sterling

Southern, Ronald E. Stinner, and C. Bruce Williams III is also greatly appreciated. Very special thanks are extended to Dr. Cavell Brownie (Department of Statistics, NC State

University) for indespensable assistance with statistical analysis of results.

Field studies would not have been possible without the active cooperation and participation of a number of golf course superintendents including David Pate (Sea Trail

Plantation), Rick Vigland and Chuck Baldwin (The Lakes Country Club), Billy Lewis (Oak

Island and Brierwood Golf Courses), Mike Claffey (Cape Fear Country Club), Sammy

Corbett (Duck Haven Golf Course), Gary ‘Cutter’ Smitter (Landfall Golf Club), Sam Osteen

(Oyster Bay Golf Links), and Terry Warlick (Olde Point Golf and Country Club.

Special thanks are also extended to Jason Cryan for assistance processing specimens in the laboratory. Student research assistants Ian Winborne and Ed Karoly and coworkers

Brian Royals and Dr. Yulu Xia are also acknowledged for long hours of sampling in the field and assistance in performing the greenhouse studies and data entry. Additional thanks are also extended to the host of Brunswick County Master Gardener volunteers that helped to collect field data and specimens for various studies.

iv TABLE OF CONTENTS

LIST OF TABLES...... vii

LIST OF FIGURES...... xi

I. FLIGHT ACTIVITY OF SCAPTERISCUS VICINUS SCUDDER

AND S. BORELLII GIGLIO-TOS (ORTHOPTERA: GRYLLOTALPIDAE)

IN SOUTHEASTERN NORTH CAROLINA...... 1

Introduction...... 2

Materials and Methods...... 5

Results...... 9

Discussion...... 22

References Cited...... 29

II. THE EFFECT OF SOIL MOISTURE ON OVIPOSITIONAL BEHAVIOR

IN THE SOUTHERN MOLE CRICKET SCAPTERISCUS BORELLII

GIGLIO-TOS (ORTHOPTERA: GRYLLOTALPIDAE)...... 41

Introduction...... 42

Materials and Methods...... 44

Results...... 48

Discussion...... 54

References Cited...... 59

v III. THE DEVELOPMENT OF SCAPTERISCUS VICINUS SCUDDER

AND S. BORELLII GIGLIO-TOS IN SOUTHEASTERN

NORTH CAROLINA...... 70

Introduction...... 71

Materials and Methods...... 73

Results...... 79

Discussion...... 90

References Cited...... 114

IV. EFFECT OF MOISTURE AND TIME OF YEAR ON MOLE

CRICKET (ORTHOPTERA: GRYLLOTALPIDAE)

SURFACE TUNNELING...... 130

Introduction...... 131

Materials and Methods...... 134

Results...... 136

Discussion...... 139

References Cited...... 143

LIST OF REFERENCES...... 150

vi List of Tables

Chapter I

1. Annual mole cricket flight trap data by site, year, species,

and sex from two sites in North Carolina for 1995, 1996,

1997, and 1998. Data are numbers (%) captured annually

in standard paired calling traps at The Lakes Country Club

(FS) and Sea Trail Plantation (ST), Brunswick Co., NC...... 33

2. Spring mole cricket flight trap data by site, year, species,

and sex from two sites in North Carolina for 1995, 1996,

1997, and 1998. Data are numbers (%) captured in

standard paired calling traps (1 Jan - 31 July) at The Lakes

Country Club (FS) and Sea Trail Plantation (ST),

Brunswick Co., NC...... 34

3. Fall mole cricket flight trap data by site, year, species, and

sex from two sites in North Carolina for 1995 and 1996.

Data are numbers (%) captured in standard paired calling

traps (1 Sept-31 Dec) at The Lakes Country Club (FS) and

Sea Trail Plantation (ST), Brunswick Co., NC...... 35

4. Calendar date estimates for mean, 25, 50 and 75%

cumulative spring flight and hatch for Scaptericsus vicinus

and S. borellii at two sites in North Carolina. (FS = The Lakes

CC, ST = Sea Trail Plantation)...... 36

vii Chapter II

1. Oviposition by S. borellii confined in chambers containing

three different soil moisture treatments in Experiment II,

summarized at day 28 and 32...... 64

2. Oviposition response of S. borellii subjected to a rapid

increase in soil moisture in Experiment III, summarized at

10 day after the flush...... 65

Chapter III

1. Pronotal midline lengths (mm) used to assign field

collected Scapteriscus vicinus and S. borellii nymphs to

size classes ( C1 - C7) equivalent to instar (adapted from

Matheny and Stackhouse 1980)...... 119

2. Percentage species composition of two species of mole

crickets from soapy water flush samples (n=10 in

1993–1996, n=12 in 1997) at two sites in North Carolina.

FS = Fox Squirrel, ST = Sea Trail...... 120

3. Julian date estimates for three percentages of peak mean

count for two size classes of S. vicinus and S. borellii in

southeastern North Carolina (1993–1997)...... 121

3a. Calendar date estimates for three percentiles of peak mean

count for two size classes of S. vicinus and S. borellii in

southeastern North Carolina (1993–1997)...... 121

viii 4. Julian date estimates (± SE) for three percentages of cumulative

abundance for two size classes of S. vicinus and

in southeastern North Carolina (1993–1997)...... 122

4a. Calendar date estimates (± SE) for three percentages of cumulative

abundance for two size classes of S. vicinus and S. borellii

in southeastern North Carolina (1993–1997)...... 122

5. Summary of mean soil degree-day accumulations

(DDEC ± SD) among sites and years at three percentages

of cumulative abundance for two size classes of S. vicinus

and S. borellii in southeastern North Carolina (1993–1997;

n = 11 for S. vicinus; n = 11 and 10 site-years for C1 and

C2 S. borellii, respectively)...... 123

6. Comparison of Julian date and degree-days (DD, base 10EC)

from 1 January for three percentages of cumulative abundance

for S. vicinus and S. borellii in southeastern NC, (1993–1997),

N = sites for date analysis, DD superscript = no. sites for DD

correlation analysis...... 124

7. Degree-day accumulations (DD, base 10EC) from 1 January

to 30 May and monthly rainfall accumulations used for

modeling date of cumulative abundance for S. vicinus C1

nymphs in southeastern NC, (1993–1997),

N = sites for date...... 125

ix 8. Comparison of effects for 30 May degree-day accumulations

and June rainfall as explanatory variables for date of 50%

cumulative abundance of S. vicinus C1 nymphs. (df = 1 for

both variables)...... 126

9. Calendar date estimates for three percentages of cumulative

oviposition for S. vicinus and S. borellii in southeastern

North Carolina (1993–1997)...... 126

Chapter IV

1. Mean mole cricket damage ratings (0–9) and mean

percentage soil moisture from bermudagrass fairways in

NC (1995, 1996 and 1997)...... 146

2. Modified analysis of covariance of the relationship

between mean damage ratings (0–9) and mean percentage

soil moisture (%SM) at 0–10.2 cm (r2 = 0.96); and in the

10.2–20.3 and 20.3–30.5 cm ranges, combined (r2 = 0.95)...... 147

x List of Figures

Chapter I

1. Cumulative spring flight counts of two species of mole

crickets at two sites in North Carolina, 1995 - 1998

(FS = The Lakes C. C., ST = Sea Trail Plantation)...... 37

2. Cumulative fall flight counts of two species of mole

crickets at two sites in North Carolina, 1995 - 1996

(FS = The Lakes C. C., ST = Sea Trail Plantation)...... 39

Chapter II

1. Comparison of the cumulative ovipositional response

of female S. borellii among the three soil moisture

treatments in Experiment II (4 blocks, n = 20)...... 66

2. Relationship between soil moisture and mean percentage

of female S. borellii ovipositing in the three moisture

treatments in Experiment II...... 67

3. Comparison of the cumulative ovipositional response

of female S. borellii in the two treatments of Experiment

III after a rapid increase in soil moisture in the flush

treatment (2 blocks, n = 35)...... 68

4. Relationship between soil moisture and matric potential

(-KPa) for the soil used in the S. borellii oviposition

experiments...... 69

xi Chapter III

1. Comparison of timing between peak mean counts of

C1 nymphs of S. vicinus and S. borellii. Data are mean

soapy water flush counts taken on golf courses in

southeastern North Carolina (1993–1997)...... 127

2. Comparison of timing between peak mean counts of

C2 nymphs of S. vicinus and S. borellii. Data are mean

soapy water flush counts taken on golf courses in

southeastern North Carolina (1993–1997)...... 128

3. Trend in correlation between date of 25 and 50%

cumulative abundance of Class 1 S. vicinus nymphs and

soil degree-day accumulations at twelve dates from golf

courses in southeastern North Carolina (n = 11 site-years,

1993–1997)...... 129

Chapter IV

1. Relationship between soil moisture and matric potential

(-KPa) for Kureb fine sand from the study site...... 148

2. Linear relationship between mean damage ratings

and Julian date in 1995, 1996 and 1997. The Lakes

Country Club, Boiling Spring Lakes, Brunswick Co.,

NC. Julian dates presented as calendar dates for illustration...... 149

xii Chapter I

Flight Activity of Scapteriscus vicinus Scudder and S. borellii Giglio-Tos (Orthoptera:

Gryllotalpidae) in southeastern North Carolina

Peter T. Hertl, Rick L. Brandenburg, and C. Bruce Williams III

______

ABSTRACT: Seasonal flights and egg-hatch of the tawny mole cricket (Scapteriscus vicinus) and southern mole cricket (S. borellii) were monitored (1995–1998) at two sites in southeastern North Carolina. Flying adults were captured using standard paired acoustic trapping stations. Egg-hatch was monitored weekly during the hatch season using a soapy water solution to bring nymphs to the soil surface. Peaks in the annual flight activity of both species occurred in the spring and fall. Spring flights represented 78.8 ± 10.0 and 88.8 ±

3.7% of the overall annual flight activity for S. vicinus and S. borellii, respectively. S. vicinus flights were earlier and had a shorter spring flight period than S. borellii in most cases. Timing of median spring flights for S. vicinus were similar to the timing of median flights reported in central Georgia. Median spring flights of S. borellii were 2–4 weeks later than reported in Georgia, suggesting that flight activity of this species may be correlated with latitude. Significant differences were noted in the timing of flight between sites and among years, but no consistent trends were evident for both species. Flight and nymph count data were used to identify when 25, 50 and 75% cumulative spring flight and hatch had occurred. These data were examined to determine if the timing of flight was correlated with date, soil degree-days, and hatch. Julian date and degree-day accumulations at 25, 50 and 75% flight show varying degrees of correlation with S. borellii hatch, and an equation predicting median hatch date is presented. The relationship between total spring flight counts, nymph population levels and damage were also examined. The number of mole crickets flying in spring was not correlated with either previous or subsequent nymph population levels, or damage levels occurring later in the season.

Key Words: Scapteriscus vicinus, S. borellii, mole crickets, acoustic calling traps, flight.

______

Mole crickets in the genus Scapteriscus are soil-dwelling pests of turfgrass throughout the southeastern United States. Two introduced species, Scapteriscus vicinus Scudder, the tawny mole cricket and S. borellii Giglio-Tos (formerly S. acletus Rehn and Hebard), the southern mole cricket are the two most important orthopteran pests throughout this region

(Brandenburg, et al. 1997). Both species have a similar one year life cycle in North

Carolina, overwintering as either large nymphs or adults with reduced surface activity during the cool winter months. Spring mating activity is associated with mate calling and flight activity. The calling and flight behavior of Scapteriscus has been well-studied in

Florida. Beginning on warm nights in early spring, males construct calling chambers at the soil surface and stridulate to attract receptive females (Nickerson et al. 1979). S. vicinus usually begins calling 10 to 20 minutes after sunset and S. borellii males follow shortly thereafter. Both species continue to attract mates with their species-specific song for 60 to

90 minutes (Forrest 1980, 1981). Females may fly more than once to the calls and some males are also attracted (Ulagaraj and Walker 1973). After mating, females oviposit in underground chambers and the eggs hatch approximately 20 days later. In North Carolina,

2 oviposition begins in April or May and the majority of eggs hatch in June and July. The feeding and tunneling of the small nymphs often goes undetected until damage caused by the larger instars becomes apparent by late July or early August (Hertl and Brandenburg

2002). Some nymphs of both species become adults in late summer and disperse to new habitats during a short fall flight period.

Flight is one of the few natural events that bring mole crickets out of the soil, and attempts have been made to exploit this behavior for both research and management purposes. Systematic monitoring of mole cricket flight activity became possible with the development of electronic recording and calling devices. Ulagaraj and Walker (1973) were the first to perform controlled studies to demonstrate that flying mole crickets are attracted to electronic reproductions of their calling songs and Ulagaraj (1975) and Ulagaraj and

Walker (1973, 1975) developed basic techniques for using acoustic callers to trap mole crickets. Since that time acoustic callers have been used to monitor mole cricket flight activity and abundance (Walker 1982, Walker and Fritz 1983), study the geographic variation in flights (Walker et al. 1983, Fowler 1987, Braman and Hudson 1993), and to collect live crickets for research (Walker 1982, 1988), or for use as fish or animal food

(Walker 1988). Fowler (1988) used electronically produced calls of Scapteriscus to attract and collect live specimens of the mole cricket parasitoid, Ormia depleta (formerly

Euphasiopterx depleta), for biological studies and Forrest (1983) used the phonotactic response to synthesized songs to differentiate Scapteriscus spp. in Puerto Rico from those found on the U.S. mainland. The potential for using acoustic calling traps to control mole crickets has also been investigated. Ngo and Beck (1982) studied flight behavior and the

3 potential for controlling S. borellii using sound traps to reduce adult populations. Calling traps have also been used to inoculate mole crickets with entomopathogenic nematodes for biological control (Parkman and Frank 1993), attract crickets into nematode-treated areas to increase the probability of infection (Parkman et al. 1993), and quantify the level of nematode infection in mole cricket populations (Parkman and Frank 1992). Additionally, the callers have been used to attract crickets into insecticide-treated areas, or away from areas of managed turfgrass.

Flight activity has a direct bearing on three factors of considerable importance in mole cricket management: mating, egg-laying, and dispersal into new or previously infested areas.

Information on the timing of mole cricket flights is critical to the understanding of these pests and may prove valuable in developing improved management programs and making management decisions. This is especially true if flight proves useful in predicting hatch.

Virtually all the previous calling trap studies in the United States have been conducted in

Florida, and only two studies, conducted in Georgia (Braman and Hudson 1993), and

Louisiana (Henne and Johnson 2001) have examined flight activity north of Florida. Due to the importance and economic impact of these pests, we pursued a program to monitor flight activity in North Carolina, near the northern extreme of their range. The objective of this study was to quantify the timing and extent of mole cricket flights in North Carolina and determine if flight data were useful in predicting events such as oviposition, egg-hatch, development, nymph population levels, or the onset and severity of damage. Further, we wanted to determine if flight activity was correlated with measurable environmental factors useful in constructing a phenologically based model for mole cricket management.

4 MATERIALS AND METHODS

Study Sites and Dates. Standard paired mole cricket trapping stations (Walker 1982) were established at two golf courses in Brunswick County, in the extreme southeastern corner of

North Carolina. The southernmost site, Sea Trail Plantation (ST, 33E 54' N, 78E 31' W ) is a complex of three 18 hole golf courses located on the mainland near Sunset Beach, just a few km from the South Carolina boarder. Species composition and the level of infestation vary somewhat among the courses, and the traps were located on the Jones Course, which had a moderate infestation of both species. The northernmost site was The Lakes Country Club

(formerly the Fox Squirrel Country Club, FS; 34E 02' N, 78E 03' W ) in Boiling Spring

Lakes, approximately 27 km SSE of Wilmington NC. The sites are separated by approximately 50 km. These sites were chosen because both had active infestations, cooperative management, and were already being sampled for nymphs as part of an ongoing study of mole cricket development. We monitored flight activity at both sites from mid-

April 1995 until 22 September 1997 at ST and until 20 August 1998 at FS. Depending on the site and year, traps were checked daily to weekly during the main part of the flight season, but were monitored less frequently during the colder months of the year.

Acoustic Traps and Calling Units. The trapping stations were constructed after those described in Parkman and Frank (1993) with electronic callers centered over a collection funnel with a 1.5 m dia opening. During the four year study we used two different models of electronic callers to attract flying mole crickets to the trapping stations. The first model used was the Mans Artificial Cricket (B. J. Mans, Mountain View, CA) and later, the Night

CallerTM (Eco-Sim, Gainesville, FL), an improved version of the Mans design. Both types

5 were developed for use in mole cricket research at the University of Florida and both are capable of simulating the mating call of either species. Each unit consists of a sound- synthesizer computer chip programmed with the song parameters of both species, an amplifier, a speaker, and a photocell to initiate calling at dusk. The units were powered with either a 115v AC/12v DC invertor or 12v DC battery. Beginning at sunset, each unit automatically broadcast the song selected at 105 dB for two hours. Calls generated for S. vicinus have a carrier frequency of 3.3 kHz and a pulse rate (wing-stroke rate) of 130 pulses/sec while the call of S. borellii have a carrier frequency of 2.7 kHz and a pulse rate of

50 pulses/sec (Ulagaraj 1976). The callers for both species utilize a 50% duty cycle.

Crickets were usually captured into ethylene glycol preservative, but live specimens for laboratory and greenhouse experiments were occasionally collected into 19 L buckets containing approximately 10 L of moist sand. All crickets were identified to species, sexed and counted.

Environmental Data. Meteorological and degree-day data were acquired from several sources. Site specific soil degree-day data were collected at both sites using model TA51-P biophenometers (Omnidata International, Inc., Logan, Utah) with a soil temperature probe located at a depth of 10.2 cm under average turf. These units were programmed to automatically calculate degree-day accumulations every 10 minutes within the range of

10–43EC, which represents the generally accepted upper and lower thermal limits of insect development. Soil degree-day accumulations (starting 1 January each year) were recorded at each sample date and visit to the site. Rainfall data were provided by the State Climate

Office of North Carolina at NC State University.

6 Nymph Sampling. Mole cricket nymphs were sampled weekly during the summer months

(end of May through July) at both sites using the soapy water flush technique of Short and

Koehler (1979). Samples were taken by applying 16 L of a 0.4% aqueous solution of dishwashing soap (Lemon Fresh Joy®) to a 1.0 m2 area of turf (delineated by a 1m x 1m

PVC frame). A sprinkling can was used to evenly distribute the irritant solution over the area and all mole cricket nymphs emerging from the soil were collected and preserved in

80% ethyl alcohol. Different fairways were sampled each week and sampling sites were randomly selected from within areas showing recent mole cricket damage, and therefore represent a judgement sample. Ten samples (n = 10) were taken each week in 1994, 1995 and 1996, and 12 samples (n = 12) per week were taken in 1997. Nymphs were later identified to species, counted, and measured using a dissecting scope equipped with an ocular micrometer. Because of the difficulty in assigning nymphs to specific instars, mid- line pronotal length was used to separate the nymphs into size classes. The smallest nymphs

(Class I) were used to approximate the newly-hatched first instar of each species. Nymphs with a midline pronotal length of 1.4–1.6 and 1.6–1.9 mm were classified as Class I nymphs for S. vicinus and S. borellii, respectively. The relative weekly counts of these Class I nymphs were used to quantify hatch for each species at each site.

Data Analysis. Flight count data were used to compile statistics on annual, spring and fall flight activity. Spring flight activity was of special interest. A spring flight termination date of 31 July (~Julian day (JD) 213) was selected, primarily to accommodate the later flight activity of S. borellii. Spring flight data were examined using PROC LIFETEST and significant differences in the timing of flight activity among years and between sites and

7 species were determined using Wilcoxon Test (Chi-square) to test equality of the flight distributions. The results of a second test, the Log-Rank Test of equality are mentioned where the results differed from the Wilcoxon analysis. Spring flight and Class I nymph counts were also used to make JD estimates of the dates at which 25, 50, and 75%

(quartiles) of each species total spring flight and hatch had occurred. Flight date estimates were made (PROC UNIVARIATE) assuming no underlying distribution. Date estimates for hatch were made (PROC LIFEREG) assuming a logistic distribution. Estimates of the mean day of flight and hatch were made in a similar manner. Julian date quartile and mean day estimates for flight and hatch were used to examine the relationship between the timing of flight and hatch for each species (PROC CORR, PROC GLM). Estimates of soil degree-day accumulations corresponding to the quartile dates and mean day of flight and hatch were used to examine the degree-day relationship between flight and hatch for each species

(PROC CORR, PROC REG). Total nymph and spring flight counts were examined to determine if there was a relationship between nymph populations and the number of flying adults. Where nymph counts were compared among years or sites, the 1997 counts were adjusted to make the 12 m2 sample counts comparable to years when 10 m2 samples were taken. Two flight-hatch scenarios were examined. Spring flight counts were compared to nymph counts of the previous year (summer) to determine if there was a relationship between nymph population levels and the number of adults flying the following spring.

Spring flight counts were also compared to nymph counts of the same year (season) to determine if there was a relationship between the number of spring-flying adults and subsequent nymph population levels (PROC GLM). The relationship between the total

8 numbers of mole crickets flying in the spring and the mean levels of mole cricket damage occurring in the fall was also examined to determine if flight could be used as a damage risk assessment tool. The timing of fall flight (flight activity after 1 September (JD 244)) between species, sites and years was examined using the same methods used to examine spring flight activity. All tests of equality, correlation, and analysis of variance (ANOVA) were performed using SAS (SAS Institute 1999).

RESULTS

Several interruptions in flight monitoring occurred due to power outages and storm damage to the calling stations. Except for one power outage at ST (15–20 June 1995) there were no other interruptions in flight monitoring the first year. In 1996, Hurricanes Bertha

(10 July) and Fran (5 September) disabled the calling traps at ST from 10 July to 20 August, and again from 5–10 September 1996. Therefore, no data were available for the 1996 spring flight study at ST after 10 July (JD 192). Traps at FS were out of service from 5–18

September in 1996. Monitoring of fall flight activity was interrupted at both sites in 1997.

Due to construction at ST, the traps were removed and the trapping study at that site ended on 22 September 1997. At FS, the S. borellii caller was non-functional from 17–21

September 1997, and neither caller worked at that site after 7 October 1997 (JD 280).

Trapping began again at FS on 1 February 1998. Spring flight data were probably unaffected as no flight activity was observed prior to this date in other years. The trapping study ended on 20 August 1998 when the trapping station at FS was destroyed by Hurricane

Bonnie.

9 Species Composition, Sex Ratio, and Numbers Captured in Flight Traps Annually.

Annual flight trap capture data are summarized by site, year, species and sex in Table 1. A total of 9411 mole crickets were captured at the two sites during the four year study. The predominant species captured was S. borellii, representing 85.7 ± 4.6% of all flight captures at both sites. Females were the predominant sex captured, comprising 92.9 ± 1.5 and 79.5 ±

1.1% of the total combined annual catch of S. vicinus and S. borellii, respectively. Data presented by Matheny et al. (1983) indicate that only 7.5% of S. vicinus and 36% of S. borellii are captured by calling traps with a 1.5 m dia capture funnel. Using these data, the adjusted annual percentage catch over all sites and years was 39.0 ± 8.4 and 61.0 ± 8.4% for

S. vicinus and S. borellii, respectively. Trapping was continued one year longer (1998) at

FS than at ST, therefore, the following comparisons between the sites are based on data collected 1995–1997. During these three years S. borellii comprised 85.6 ± 5.4% of the overall combined catch at both sites (83.2 ± 10.5 and 87.9 ± 5.6% at FS and ST, respectively). Females represented 80.0 ± 1.2% of the overall combined total catch for S. borellii (78.2 ± 2.0% and 81.8 ± 0.1% at FS and ST, respectively). Females represented

93.9 ± 1.4% of the total combined catch for S. vicinus at both sites (93.1 ± 1.5% and 94.7 ±

2.7% at FS and ST, respectively).

There was significant annual variation in species composition and numbers flying among years at the two sites, however, few conclusions or trends can be drawn from the data.

Comparison of S. borellii annual capture numbers between sites or among years is somewhat difficult because in 1996 the traps at ST were damaged while S. borellii spring flights were still occurring. This undoubtedly resulted in lower catch numbers for this

10 species in 1996. Comparison of count data for S. vicinus is less problematic because the flight activity of this species had already ended before the ST traps were damaged (JD192).

Although annual catch counts were also affected by trap damage (FS) or removal (ST) during the fall flight season of 1997, the greatest annual trap captures of S. borellii were observed at both sites in 1997. The greatest number of flying S. vicinus were captured during 1995 at FS and 1997 at ST. Both sites had extremely low flight activity for S. vicinus in 1996.

Species Composition, Sex Ratio, and Numbers Captured During Spring Flights.

Generally there were two peaks in annual flight activity for each species. Major flight activity for both species occurred in the spring, and much reduced flight activity was recorded in the fall. Spring flight data (1 Jan – 31 July) are summarized by site, year, and species in Table 2. Data on spring flight were available each year of the study and suffered far less from the weather-related problems that plagued the fall flight study. S. borellii was the predominate species captured, representing 86.4 ± 4.1% of all spring-captured mole crickets for all years at both sites. Females were the predominate sex captured during the spring flight period for all years at both sites, representing 92.4 ± 1.6 and 79.5 ± 1.3% of all spring-captured S. vicinus and S. borellii, respectively. Spring flights represented the majority (86.9 ± 3.6%) of the total annual flight activity recorded in 1995 and 1996, the only years that annual capture records were complete enough for comparison. During these two years, spring flight accounted for 78.8 ± 10.0 and 88.8 ± 3.7% of the total annual flight catch for S. vicinus and S. borellii, respectively, with S. borellii representing 91.3 ± 1.2 and 86.3 ±

5.3% of the catch at FS and ST, respectively. Comparison of spring catch between sites was

11 only possible in 1995, 1996, and 1997. During these three years S. borellii represented 85.3

± 8.9 and 87.5 ± 5.9% of the spring catch at FS and ST respectively, with an overall percentage of 86.4 ± 4.8% over both sites.

Actual percentage catch for each species is reported by year and site in Table 2. Catch adjusted according to Matheny (1983) indicate that the relative percentage of mole crickets attracted at FS were 30.7, 89.7, 65.3, and 57.0% S. borellii for 1995, 1996, 1997, and 1998, respectively, with an overall mean of 60.7 ± 12.2%. Adjusted catch at FS for 1995–1997 was 70.7 ± 9.8% S. borellii. Adjusted catch at ST was 62.2, 87.3, and 40.7% S. borellii for

1995, 1996, and 1997, respectively, with an overall mean of 63.4 ± 13.5%. Adjusted catch over all sites and years was 61.9 ± 8.3% S. borellii. Adjusted catch over both sites for 1995

–1997 was 62.7 ± 9.7% S. borellii.

Spring flight data were further analyzed to determine the relative timing of flights between species, sites, and among years. Because of the relationship between spring flights and reproduction, the timing of flight and hatch was also examined for each species.

Species Composition, Sex Ratio, and Numbers Captured During Fall Flights. Due to various interruptions caused by storms, construction, or equipment failures, fall flight data were not available for either site in 1997 or 1998. Although Hurricane Fran resulted in the loss of six trap-nights at ST (5–10 Sept ) and 14 trap-nights (5–18 Sept) at FS in 1996, the eight additional nights of trapping at ST produced only one female S. vicinus, and three female S. borellii, and did not seriously compromise comparison of the two sites. Therefore, comparisons of fall flight data among years or sites is based solely on data from 1995 and

1996 (Table 3). During these two years, fall flights represented only 12.4 ± 2.8% of the

12 total annual flight captures for both species (20.9 ± 7.0% and 10.5 ± 2.8% for S. vicinus and

S. borellii, respectively). Fall flights captures of S. borellii represented 79.2 ± 16.2% of the overall combined fall catch at the two sites (60.9 ± 29.8 and 97.5 ± 2.5% at FS and ST, respectively). Female S. borellii comprised 80.5 ± 1.8% of the total fall catch for that species at both sites. Although similar total numbers of mole crickets were captured during fall flights at FS, there were significant differences in species composition between these two years. In 1995, 68.9% of the fall flight catch were S. vicinus, but this species represented only 9.3% of the fall flight captures in 1996. Conversely, there was no fall flight activity recorded for S. vicinus at ST in 1995, and only six female of that species flew at that site in the fall of 1996. Excluding the zero fall catch at ST in 1996, fall flight represented 27.9 ± 0.8% of the combined annual catch for S. vicinus at the two sites.

Females comprised 98.4 ± 1.6% of the non-zero S. vicinus fall flight catch.

Fall flight data adjusted according to the proportions reported by Matheny (1983) indicate that S. borellii represented 8.6 and 67.1% of the fall catch at FS for 1995 and 1996, respectively, with an overall mean of 37.8 ± 29.2%. Similar calculations show that S. borellii comprised 100.0 and 80.1% of mole crickets attracted at ST in the fall of 1995 and

1996, respectively, with an overall value of 90.1 ± 9.9%. According to these calculations S. borellii comprised 63.9 ± 19.7% of all fall-attracted mole crickets over both years and sites.

Relative Timing of Spring Flights Between Species. Timing of spring flights is presented graphically in Figure 1. S. vicinus flew significantly earlier than S. borellii at both sites and all years except 1995. In 1995, a three night catch (6 May) at FS produced the largest single

S. vicinus flight count observed at either site in any year, concurrent with one of the largest

13 (and earliest) S. borellii flight counts of the study. This single flight represented 57.7 and

27.6% of the total spring flight activity at FS for S. vicinus and S. borellii, respectively, and resulted in both species having the same median day of spring flight (Table 4), and was the only case where there was no significant difference in the timing of spring flights between the two species. Over all other years and sites there was a mean difference of 47.4 ± 6.0 days in mean day of flight between species, and a mean difference among the three flight quartiles of each species of 52.4 ± 4.5 days. Flight activity of S. vicinus was significantly earlier than S. borellii at FS in 1996 (Chi sq = 116.1483, df = 1, P < 0.0001), 1997 (Chi sq =

1276.3894, df = 1, P < 0.0001), and 1998 (Chi sq = 785.6093, df = 1, P < 0.0001). Date estimates for these three years indicate that flights of S. vicinus were 51.3 ± 3.7, 55.2 ± 12.0,

62.3 ± 14.2 days earlier for 25, 50, and 75% flight (quartiles), respectively, with difference of 48.0 ± 7.7 days in mean day of flight. At ST, S. vicinus flights were significantly earlier than the flights of S. borellii all three years of the study (Chi sq = 540.3 (1995), 344.6

(1996), 1686.9 (1997), df = 1, P < 0.0001). Date estimates for the 25, 50 and 75% flight quartiles indicate that S. vicinus flights were an average of 42.3 ± 12.1, 49.0 ± 15.1, 54.0 ±

13.6 days earlier, respectively, than the flights of S. borellii, with a mean difference of 46.9

± 10.9 days in mean day of flight.

Relative Timing of Spring Flights Among Years. Both Log-Rank and Wilcoxon tests of equality show highly significant differences in the timing of spring flight among years (P <

0.0001) for each species at both sites. The two tests of equality are somewhat different. The

Log-Rank test is generally better at detecting differences between distributions when the distributions cross. The Wilcoxon test is better at detecting shifts in the overall timing

14 between distributions. Shifts in the overall flight distribution were deemed as more meaningful, therefore the results of the Wilcoxon test are reported unless otherwise noted.

Although the timing between sites differed some years, the timing of S. borellii spring flights show a similar pattern among years at both sites (Table 4). S. borellii flight activity at both sites was significantly earlier in 1995 than in all other years. Flight activity at both sites was significantly earlier in 1996 than in 1997 (Chi sq = 109.9648, df = 1, P < 0.0001;

Chi sq = 46.6669, df = 1, P < 0.0001, for FS and ST, respectively); and at FS, flights in 1997 were significantly earlier than in 1998 (Chi sq = 13.2534, df = 1, P < 0.0003). Because sampling at ST in 1997 was terminated before the usual end date, it was expected that the distribution of S. borellii spring flight might appear earlier than other years. However, as both Wilcoxon and Log-Rank tests indicate that 1997 was the latest year at ST, it appears that the early end date had a minimal effect. In contrast, S. vicinus flight activity at both sites in 1997 was significantly earlier than in any other year (Table 4). The timing of flights were not significantly different in 1995 and 1996 at ST (Chi sq = 0.0055, df = 1, P <

0.9410), however, at FS, S. vicinus flights in 1996 were significantly later than in 1995 (Chi sq = 6.3195, df = 1, P < 0.0119). It should be noted that the Log-Rank test results indicate that S. vicinus flights at FS were not different in 1995 and 1996 (Chi sq = 2.5870, df = 1, P <

0.1077). There was no significant difference between the timing of flights in 1996 and

1998 (Chi sq = 3.1078, df = 1, P < 0.0779).

The earliest spring flight date for both species at either site was observed the first week of March in 1997. Date of first flight did not seem to have any relationship with the overall timing of flight for S. borellii, however, both sites had significantly earlier S. vicinus flight

15 activity in 1997, suggesting a relationship for that species.

Relative Timing of Spring Flights Between Sites. Comparison of spring flight data between the two sites was only possible in 1995, 1996, and 1997. Analysis indicates there were significant differences in the timing of spring flights between the two sites (Table 5), however, no consistent trends are apparent in the timing of flight of either species. Results of the Wilcoxon test indicate that the timing of S. borellii flight activity was earlier at FS in

1995 (Chi sq = 299.5446, df = 1, P < 0.0001), earlier at ST in 1996 (Chi sq = 12.7335, df =

1, P < 0.0004), but was similar in 1997 (Chi sq = 1.3928, df = 1, P < 0.2379). Flight activity of S. vicinus was earlier at ST in 1995 (Chi sq = 89.6670, df = 1, P < 0.1001) and 1997 (Chi sq = 79.3559, df = 1, P < 0.1001), but similar in 1996 (Chi sq = 0.4054, df = 1, P < 0.5243).

As seen before, the 1995 data from FS and the switching of early / late status between quartile date estimates for the two sites provide additional confusion in interpreting the results. Flights of S. borellii were earliest at both sites in 1995, and mean day of flight for this species was 15.0 d earlier at FS than at ST. Although analysis indicates that flights were earlier at ST in 1996, the sites switch early / late status for each quartile. Additionally, data from both 1996 and 1997 show that the mean day of spring flight for S. borellii was only 2.1 ± 0.0 d different between sites. Therefore, there appears to be no biologically significant difference in the timing of flight between the sites for S. borellii in 1996 and

1997. Conversely, S. vicinus flights were earlier at both sites in 1997, however, ST was 11 and 9 days earlier than FS for the 25th and 50th quartiles (respectively), but 13 days later than FS for the 75th quartile estimate. The mean day of flight is 5.4 days earlier at ST than

FS in 1997, but this could be the result of the early end date at ST. Although ST was

16 numerically earlier for all quartile date estimates in 1996, and had a mean day of flight was

5.1 days earlier at ST in 1996, both the log-rank and Wilcoxon tests of equality indicate no significant differences in S. vicinus flight between sites.

Species Composition of Nymph Samples. The results of nymph sampling indicate there were significant differences in nymph populations among years and between sites. Nymph data for 1994 are included only for the analysis comparing total nymph counts with total flight counts the following spring. Total nymph counts from 1995–1997 are also included in this study, but weekly count data for these years were used to examine the timing relationship between flight and hatch. Comparable sampling at FS indicate declining populations of S. vicinus nymphs for the years reported. S. vicinus comprised 74.4% (821) of the nymphs collected at FS in 1994. Sampling at FS in 1995 produced almost equal numbers of S. vicinus and S. borellii nymphs (702 (51.2%) and 670 (48.8%), respectively), however, much lower numbers of S. vicinus were recovered in subsequent years. S. vicinus nymphs represented only 11.7 and 12.6% of the sample totals in 1996 and 1997, respectively. Similar numbers of S. borellii were collected each year in 1995–1997. The small number of S. vicinus Class I nymphs in the samples were insufficient to characterize hatch for that species at FS in 1996. Therefore, flight and hatch data for S. vicinus could only be compared in 1995 and 1997, but S. borellii flight and hatch could be compared at FS all three years. Conversely, few nymphs of S. borellii were found in the ST samples any year, and only S. vicinus hatch could be characterized and compared with flight timing at that site.

Relationship Between Timing of Spring Flight and Egg-Hatch by Julian Date. Julian

17 date estimates for mean date and the date of 25, 50 and 75% (quartiles) of spring flight and hatch were examined for both species. As previously noted, low numbers of Class I nymphs made it impossible to quantify hatch for both species at both sites each year, limiting the data available for comparison. Correlation analysis for S. vicinus indicate no significant relationship between means or any date quartiles for flight and hatch. Correlation analysis for S. borellii indicate a significant relationship between the date of 25% flight and both

50% (n =3, Pearson Correlation Coefficient (PCC) = 0.99883, P = 0.0308) and mean hatch date (n =3, PCC = 0.99766, P = 0.0435). Date of 50% flight was correlated with both 50% hatch (n =3, PCC = 0.99750, P = 0.0451) and mean hatch date (n =3, PCC = 0.99871, P =

0.0323). The date of 25% hatch was correlated with both mean day of flight (n =3, PCC =

0.99948, P = 0.0205) and 75% flight (n =3, PCC = 0.99979, P = 0.0130). Regression analysis of equivalent date quartiles for S. borellii reveals a significant relationship between median flight and hatch dates (df = 1, t = 14.10, P = 0.0451, R2 = 0.99), and relates these dates by the model Y = 158 + 0.21646 X (where Y = Julian date of median hatch, X = Julian date of median spring flight). However, this model is only based upon the comparison of flight and hatch data for three years at one site. Therefore, caution should be used in interpreting or applying these results. Another way to view this relationship is in terms of the time between median flight and hatch. Comparison of dates for S. borellii indicate a 2–4 week difference in median flight date among years, however, date of median hatch varied only 11 days for the same years. Time between median flight and hatch varied from 59–20 days with the period between the two events decreasing with lateness of flight date (Table

4).

18 Soil Degree-Day Relationship Between Spring Flight and Egg-Hatch. Site-specific soil degree-day estimates corresponding to the Julian dates for mean day and 25, 50 and 75%

(quartiles) of spring flight and hatch were examined using the same methods used to examine Julian date. Again, low numbers of Class I nymphs made it impossible to quantify hatch for both species at both sites each year, limiting the data available for comparison. As with the analysis of date, correlation analysis for S. vicinus indicate no significant relationship between degree-day accumulations at mean or any quartile of flight and hatch.

Correlation analysis for S. borellii indicate a significant relationship between soil degree- days at 50% flight with those at 25% hatch (n =3, PCC = 0.99905, P = 0.0277). Degree- day accumulations at mean day of flight were similarly related to those at 25% hatch (n =3,

PCC = 0.99911, P = 0.0268). However, regression failed to reveal any significant relationship between degree-day accumulations for similar quartiles of flight or hatch for either species at any date tested. The differences in the biology of these two species required they be examined separately, however, because of the limited number of dates available for species-specific comparison, an attempt was made to correlate both date and degree-day accumulations for both species together. There were significant correlations at all mean and quartile dates for flight and hatch, and similar correlations for degree-day accumulations at those dates. The strength of these associations appears to be mainly due to the data for S. borellii.

Relationship Between Spring Flight Counts and Nymph Population Levels. The relationship between spring flight captures and nymph populations occurring later in the season was examined to determine if flight data were useful in predicting future infestation

19 levels. The total number of each species captured during spring flights were compared with the total number of conspecific nymphs collected at the same site each year using correlation analysis. The data reveal no correlation between the numbers of adults that flew and nymph sample counts later in the season (df = 10; P = 0.9603) indicating that flight numbers were not a good predictor of subsequent nymph population levels. Analysis using flight counts adjusted according to capture probabilities presented by Matheny et al. (1983) produced similar results (df = 10; P = 0.3176).

The relationship between the previous year’s nymph population and the number of adults flying the following spring was examined to determine if nymph counts could be used to estimate spring flight numbers. The total number of nymphs for each species were compared to the total number of adults that flew the following spring using correlation analysis. No significant relationship was found between nymph counts and flight counts for either species (df = 12; P = 0.8183) indicating that nymph counts were not a good predictor of the level of flight activity the following spring. Analysis using flight counts adjusted according to capture probabilities presented by Matheny et al. (1983) produced similar results (df = 12; P = 0.3967).

Relationship Between Spring Flight Counts and Subsequent Damage Levels. Flight counts and damage levels were examined to determine if flight counts could be used in damage risk assessment. The total spring flight count of each species, as well as the combined total spring flight count for both species were compared to mean damage levels previous reported at FS in 1995, 1996 and 1997 (Hertl and Brandenburg 2002) using correlation analysis. Mole crickets spring flight counts were not correlated with subsequent

20 mole cricket damage levels for S. vicinus (df = 1; P = 0.5474), S. borellii (df = 1; P =

0.1528), or the combined counts of both species (df = 1; P = 0.3227). Analysis using flight counts adjusted according to capture probabilities presented by Matheny et al. (1983) produced similar results for S. vicinus, (df = 1; P = 0.5475), S. borellii (df = 1; P = 0.1528), and the combined counts of both species (df = 1; P = 0.9334).

Relative Timing of Fall Flights Between Species, Sites and Years. Timing of fall flight could only be examined in 1995 and 1996. Numbers captured are reported in Table 3 and cumulative fall flight is presented graphically in Figure 2. No S. vicinus were captured at

ST in fall 1995 and only 6 were captured in 1996. Only 13 S. vicinus were captured in fall at FS in 1995. Estimates of flight timing were not compared where less than 15 captures were recorded. Therefore S. vicinus timing could not be compared between sites or years.

At FS, S. borellii flights were earlier than S. vicinus flights in 1995 (Chi sq = 20.1465, df =

1, P < 0.0001). Median and mean flight dates indicate that S. borellii flew 4–6 days earlier than S. vicinus. The dates of 25, 50 and 75% cumulative fall flight for S. vicinus was 14, 28

October and 3 November, and 4, 24 and 25 October for S. borellii. Timing of fall flight for

S. borellii was not significantly different in 1995 and 1996 at either FS (Chi sq = 0.0359, df

= 1, P = 0.8497) or ST (Chi sq = 2.3485, df = 1, P = 0.1254). Fall flights of S. borellii were significantly earlier at ST than FS in 1996 (Chi sq = 9.8822, df = 1, P = 0.0017) with mean and median flight dates indicating S. borellii were 4–7 days earlier at ST than FS. Fall flight activity of both species began in mid-September with the majority of all fall flight occurring in October. Over both sites and years, 75% cumulative flight of both species had occurred as early as 14 October and as late as 3 November.

21 DISCUSSION

Regardless of the year, site or relative population levels indicated by nymph sampling, the overwhelming majority of mole crickets captured in our North Carolina flight study were S. borellii. This may be the result of actual differences in flight activity, local placement of the traps, or a variety of other factors. One such factor to consider is that the two species are not captured equally by the traps. Matheny et al. (1982) demonstrated that although 36% of S. borellii attracted will land within the 1.5 m dia opening of a standard calling trap, only 7.5% of S. vicinus attracted will be captured within the same area.

Therefore, even if equal numbers of both species flew to the traps, almost five times as many S. borellii would be caught. Additionally, individual S. borellii are known to fly more than once in the spring (Ngo and Beck 1982), increasing the probability of capture for that species. Because the study sites were located close to the northern edge of both species current range, another possibility is suggested by the findings of Braman and Hudson

(1993). They reported that catch composition in Georgia exhibited significant clinal variation with the number of S. vicinus decreasing with increasing latitude. Although there was insufficient latitudinal difference to detect such a trend in our study, our sites are even further North and we captured significantly more S. borellii than S. vicinus. However, if there is such a trend, it is clearly a phenomenon related to flight tendency rather than population. Nymph sampling at both sites indicate that S. vicinus were present, but the proportions of flying adults and conspecific nymphs were entirely unrelated by year or site.

For example, although the nymph sampling data over all years indicate the population at ST was over 98% S. vicinus, over 87% of mole crickets captured in the flight traps were S.

22 borellii. The large numbers of flying S. borellii adults may have come from more distant sites or adjacent non-turf habitats (pond edges, woods or roughs) not sampled for nymphs, however, this does not explain the low number of S. vicinus flight captures at the same site where high populations of S. vicinus nymphs were documented. Walker et al. (1983) stated that local placement of the trapping stations influences the numbers of mole crickets caught, however, we are unsure how this affected the trapping results obtained here. Neither actual flight capture numbers nor flight numbers adjusted according to Matheny (1983) seem to represent the abundant S. vicinus population at this site.

It would seem obvious that there should be some relationship between nymph and adult population levels at the same site, however, no relationship between the two was found in our study. Flight numbers did not predict subsequent nymph numbers, nor did nymph numbers predict subsequent flight numbers. This suggests that flight captures are not a valid measure of site-specific population levels. Nymph counts were the result of fairway sampling at each site, and must be considered a more valid assay of fairway population levels and composition than flight. Nymph movements are limited to the site but mole cricket adults are known to fly at least 3.8 km Walker and Fritz (1983) and can probably fly as far as 15 km under ideal conditions (Ulagaraj 1975). Although the proximate causes of flight are not yet fully understood, and it is likely that flight tendency is influenced more by climatic factors than existing population levels. Additional factors influencing the relationship between nymph and flight numbers include the myriad of environmental and biological factors affecting survival of the eggs and mortality of the nymphs and adults.

Females were the predominant sex captured for both species in all years at both sites.

23 This is in agreement with the findings of other flight studies (Ulagaraj 1975), and is to be expected because the insects are responding to the calls made by the male of the species.

Additionally, the trap designs used are know to catch a higher proportion of females because males tend to land farther away from the sound source than the females (Matheny et al.

1983). This observation was supported in our study by the larger proportion of males captured in linear pitfall traps located adjacent to the traps (unpublished data).

Relative Timing of Spring Flights Between Species. Differences in the relative timing of flight between the species were similar to that observed in previous studies (Fowler 1987,

Ulagaraj 1975) with S. vicinus median flight date generally 4–11 weeks earlier than S. borellii. However, this was not always the case, as seen at FS in 1995. The patterns of activity observed at FS that year proved to be one of the most confusing factors in the timing study. Unfortunately, this was the first year of the study, and trapping was begun after the first flights of S. vicinus may have started. Shortly thereafter, the largest flight of S. vicinus observed in the study occurred, coinciding with a large flight of S. borellii occurring earlier than any other year in the study. This was the only site and year where no significant difference was found between the timing of flight for the two species, however, examination of climatic data for that period may provide an explanation. Data from nearby Southport indicate 1.8 and 0.2 cm of rainfall occurred on 3 and 5 May, respectively. These rain events occurred during this high catch period and followed the driest April (total monthly rainfall

1.0 cm) observed during the study. Rainfall after a period of drought has been related to the timing of flight by several authors (Hayslip 1943, Ulagaraj 1975) and may explain our results as well.

24 Relative Timing of Spring Flight Between Sites. There were significant differences in the timing of spring flights between the two sites, but no consistent trends were are apparent.

The 1995 flight data also confuses the analysis of flight timing between the sites, and this was the only year where flights of S. borellii were earlier at FS than ST. S. vicinus spring flights at ST were earlier two out of three years, however, there was no consistent pattern of parallel earliness at ST for S. borellii in the same years. Spring flight distributions of S. borellii were significantly earlier at ST only one out of three years. Earlier flight distributions might have been expected at ST, as it was the southernmost site, however, our data do not indicate such a trend. The small difference in latitude between the two sites was likely not enough to show a latitudinal effect. Additionally, the majority of S. vicinus overwinter as adults and therefore this species may have more flexibility in when they fly in the spring.

Relative Timing of Spring Flights Among Years. There were highly significant differences in the timing of spring flight among years for each species at both sites. In general, the relative differences in timing among years followed a similar pattern for each species at both sites. However, as mentioned previously, there was no trend or similarity in the timing of both species among years. Although soil moisture and a variety of other factors can vary between sites, it is curious that similar timing is evident for either species at both sites among years, but no trend in timing is evident for both species among years. The differences in the timing of flights and factors influencing the initiation of flight for each species may be different enough to preclude a pattern of similarity of flight behavior among years.

25 Although comparable data from other states are not available for the same years as the

NC study, median spring flights of S. vicinus occurred 31 March - 6 May which agrees closely with the timing of median flights in central Georgia reported by Braman and Hudson

(1993). Median spring flights of S. borellii occurred 6 May - 25 June which is 2–4 weeks later than reported by Braman and Hudson (1993). The later timing of S. borellii flight in

NC agrees with the expected variation in flight timing as a function of latitude (Fowler

1987), however, the similar timing of S. vicinus flights do not. Although Braman and

Hudson (1993) noted that flight activity for S. vicinus at their southernmost sites preceded that at more northern sites, examination of their data indicates that this was not the case for

S. borellii. This suggests that median flight date may not adequately quantify differences in flight timing for comparative purposes, or that each species responds to climatic seasonality differently.

Relationship between Timing of Spring Flights and Egg-Hatch by Date and Degree-

Days. A relationship between date of flight and date of hatch was found for S. borellii but not S. vicinus. An equation is presented to predict the date of median hatch using median flight date, however, this model is based on only three years data at one site. Median hatch date for S. vicinus was the same for both years at FS, varied less than three days during the three years at ST, and less than seven days between sites in 1995 and 1997. Median hatch date for S. borellii varied approximately 11 days among years at FS, and are highly correlated with the regression line.

As seen with date, soil degree-day accumulations for flight and hatch were correlated for

S. borellii but not S. vicinus. This is not surprising, as date and degree-day accumulations

26 are closely correlated. However, no similar degree-day model predicting hatch from flight was forthcoming, and may be partially due to the variability exhibited in the degree-day data. Although degree-days may be related to the physiology of development, flight is a behavioral phenomenon related to air temperature, wind, rainfall and a host of other factors unrelated to degree-days. Therefore, the relationship between flight and hatch is better quantified by date than degree-days.

Unfortunately, soil degree-days did not explain the relationship between median flight and hatch for S. borellii. Time between median flight and hatch varied substantially with earlier flight dates having a longer time period until median hatch than median flights occurring later in spring. Differences in heat accumulation at different points in the season would logically explain the longer time required when flights are early, however, degree- day accumulations between these dates were not similar, and do not explain the relationship.

Spring Flight Counts and Damage Levels. Spring flight numbers were not correlated with damage. The number of adults flying in the spring did not explain the differences in the mean level of damage among years occurring at FS later in the year. As discussed previously, spring flight numbers were not a good predictor of subsequent nymph population levels, and therefore are not a good predictor of subsequent damage levels.

Timing of Fall Flight. The reasons why mole crickets fly in the fall are unknown, but are believed to be related to dispersal. Some differences in timing of fall flights were noted, however, the limited data available make it difficult to draw definitive conclusions.

However, our data indicate that the main fall activity of both species occurs in October with

50–75% cumulative fall flight generally occurring in the last half of the month.

27 CONCLUSION

Spring flights of both species were later in NC than those reported previously in more southernly locations. Although the results of our study indicate that flight data may be useful for predicting hatch by date for at least one species, and resulted in an equation predicting median hatch from flight date, we did not have enough years of comparative flight and hatch data to fully resolve the relationship into a predictive function. Where such flight data are available, our flight-hatch model can be considered as a starting point for predicting the timing of hatch until more comparative studies are performed. There may be some merit to this approach, but predicting hatch by calendar date or degree-days is probably a more practical method for timing scouting or insecticide applications. The calling units still remain a valid research tool and may still prove useful in attracting the flying adults into areas that have been treated with insecticides, attracting them away from areas of managed turf, or for disseminating biological control agents such as Beauveria or entomopathogenic nematodes.

ACKNOWLEDGMENTS

The authors would like to extend their sincere thanks to superintendents David Pate (Sea

Trail Plantation), Rick Vigland and Chuck Baldwin (The Lakes Country Club) for their active participation in the study as well as a host of Brunswick County Master Gardener volunteers that helped to check and collect specimens from the traps. Very special thanks are extended to Dr. Cavell Brownie (Department of Statistics, NC State University) for indespensable assistance with statistical analysis of results. This work was funded in part by a grant from the United States Golf Association Greens Section Research.

28 REFERENCES CITED

Alexander, R. D. 1975. Natural selection and specialized chorusing behavior in insects.

Pages 35–77 in Insects, science and society. Academic Press, New York. 284 pp.

Bennet-Clark, H. C. 1970. The mechanism and efficiency of sound production in mole crickets. J. Exp. Biol. 52: 619–652.

Braman, S. K., and W. G. Hudson. 1993. Patterns of flight activity of pest mole crickets in Georgia. Intl. Turf. Soc. Res. J. 7:382–384.

Brandenburg, R. L., P. T. Hertl and M. G. Villani. 1997. Integration and adoption of a mole cricket management program in North Carolina, USA. Intl. Turf. Soc. Res. J. 8:

973–79.

Forrest, T. G. 1980. Phonotaxis in mole crickets: its reproductive significance. Fla.

Entomol. 63(1): 45–53.

Forrest, T. G. 1981. Acoustic behavior, phonotaxis, and mate choice in two species of mole crickets (Gryllotalpidae: Scapteriscus). MS. thesis, Univ. of Florida, Gainesville. 72 p.

Forrest, T. G. 1991. Power output and efficiency of sound production by crickets. Behav.

Ecol. 2(4): 327–338.

29 Fowler, H. G. 1987. Geographic variation in flight periodicity of new world mole crickets.

J. Interdiscipl. Cycle Res. 18(4): 283–286.

Hayslip, N. C. 1943. Notes on biological studies of mole crickets at Plant City, Florida.

Fla. Entomol. 26(3): 33–46.

Henne, D. C. and S. J. Johnson. 2001. Seasonal distribution and parasitism of

Scapteriscus spp. (Orthoptera: Gryllotalpidae) in southeastern Louisiana. Fla. Entomol.

84(2): 209–214.

Hertl, P. T., and R. L. Brandenburg. 2002. Effect of soil moisture and time of year on mole cricket Orthoptera: Gryllotalpidae) surface tunneling. Environ. Entomol. 31(3):

476–481.

Matheny, E. L., Jr., R. L. Kepner and K. M. Portier. 1983. Landing distribution and density of two sound-attracted mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus).

Ann. Entomol. Soc. Am. 76(2): 278–281.

Ngo Dong, and H. W. Beck. 1982. Mark-release of sound-attracted mole crickets: flight behavior and implications for control. Fla. Ent. 65(4): 531–538.

Nevo, E., and A. Blondheim. 1972. Acoustic isolation in the speciation of mole crickets.

30 Ann. Entomol. Soc. Am. 65: 980–981.

Parkman, J. P., and J. H. Frank. 1992. Infection of sound-trapped mole crickets,

Scapteriscus spp., by Steinernema scapterisci. Fla. Ent. 75(1): 163–165.

Parkman, J. P., and J. H. Frank. 1993. Use of a sound trap to inoculate Steinernema scapterisci (Rhabditida: Steinernematidae) into pest mole cricket populations (Orthoptera:

Gryllotalpidae). Fla. Ent. 76(1): 75–82.

Parkman, J. P., W. G. Hudson, J. H. Frank, K. B. Nguyen, and G. C. Smart, Jr. 1993.

Establishment and persistence of Steinernema scapterisci (Rhabditida: Steinernematidae) in field populations of Scapteriscus spp. mole cricket (Orthoptera: Gryllotalpidae). J Entomol.

Sci. 28(2): 182–190.

Ulagaraj, S. M. 1975. Mole crickets: ecology, behavior, and dispersal flight (Orthoptera:

Gryllotalpidae: Scapteriscus). Environ. Entomol. 4(2):265–273.

Ulagaraj, S. M. 1976. Sound production in mole crickets (Orthoptera: Gryllotalpidae:

Scapteriscus). Ann. Entomol. Soc. Am. 69(2): 299–306.

Ulagaraj, S. M., and T. J Walker. 1973. Phonotaxis of crickets in flight: attraction of male and female crickets to male calling songs. Science. 182: 1278–1279.

31 Ulagaraj, S. M., and T. J Walker. 1975. Responses of flying mole crickets to three parameters of synthetic songs broadcast outdoors. Nature. 253: 530–532.

Walker, T. J. 1982. Sound traps for sampling mole cricket flights (Orthoptera:

Gryllotalpidae: Scapteriscus). Fla. Entomol. 65(1): 105–109.

Walker, T. J. 1988. Acoustic traps for agriculturally important insects. Fla. Entomol.

7(4): 484–492.

Walker, T. J., and G. N. Fritz. 1983. Migratory and local flight in mole crickets,

Scapteriscus spp. (Gryllotalpidae). Environ. Entomol. 12(3): 953–958.

Walker, T. J., J. A. Reinert, and D. J. Schuster. 1983. Geographical variation in flights of the mole cricket, Scapteriscus spp. (Orthoptera: Gryllotalpidae). Ann. Entomol. Soc. Am.

76(3): 507–517.

32 Combined

S. borellii paired calling traps at The Lakes Country Club (FS) and ) (%) Females No. Males (%) No. Total (%) Total

S. vicinus Annual mole cricket flight trap data by site, year, species, and sex from two sites in North Carolina for 1995, 1996, cricket flight trap data by site, year, species, and sex from Annual mole FS 1995FS 1996FS 427 (90.5) 1997FS 43 (95.6) 1998 198 (93.4) 45 (9.5) 2 (4.4) 114 (87.0) 472 (37.4) 14 (6.6) 45 (2.9) 17 (13.0) 589 (74.5) 212 (10.0) 131 (13.5) 1172 (79.0) 202 (25.5) 1547 (81.1) 643 (76.5) 791 (62.6) 312 (21.0) 361 (18.9) 197 (23.5) 1484 (97.1) 1263 1908 (90.0) 840 (86.5) 1529 2120 971 ST 1995ST 1996ST 118 (91.5) 1997 21 (100.0)ST 1998 346 (92.5) 11 (8.5) 0 (0.0) ----- 129 (10.4) 28 (7.5) 21 (3.3) 910 (81.7) 374 (22.6) 500 (81.8) ----- 1048 (81.9) 204 (18.3) 111 (18.2) 1114 (89.6) 231 (18.1) 611 (96.7) ----- 1279 (77.0) 1273 632 1653 ------Site Year (%) No. Females Males (%) No. Total (% Table 1. 1997, and 1998. Data are numbers (%) captured annually in standard 1997, and 1998. Data are numbers Sea Trail Plantation (ST), Brunswick Co., NC. 33 Combined

S. borellii ) (%) Females No. Males (%) No. Total (%) Total

S. vicinus Spring mole cricket flight trap data by site, year, species, and sex from two sites in North Carolina for 1995, 1996, 1997, cricket flight trap data by site, year, species, and sex from Spring mole FS 1995FS 1996FS 304 (88.6) 1997FS 30 (93.8) 1998 195 (93.3) 39 (11.4) 114 (87.0) 343 (31.9) 2 (6.2) 14 (6.7) 537 (73.5) 17 (13.0) 32 (2.3) 209 (10.0) 131 (13.6) 194 (26.5) 1056 (79.0) 1531 (81.0) 640 (76.6) 731 (68.1) 281 (21.0) 358 (19.0) 1074 195 (23.4) 1337 (97.7) 1889 (90.0) 835 (86.4) 1369 2098 966 ST 1995ST 1996ST 118 (91.5) 1997ST 15 (100.0) 1998 346 (92.5) 11 (8.5) 0 (0.0) ----- 129 (11.2) 28 (7.5) 15 (2.9) 839 (82.3) 374 (23.3) 408 (82.4) ----- 181 (17.7) 1010 (81.9) 87 (17.7) 1020 (88.8) 223 (18.1) 495 (97.1) ----- 1149 1233 (76.7) 510 1607 ------Site Year (%) No. Females Males (%) No. Total (% Table 2. and 1998. Data are numbers (%) captured in standard paired calling traps (1 Jan - 31 July) at The Lakes Country Club (FS) and and 1998. Data are numbers Sea Trail Plantation (ST), Brunswick Co., NC.

34 on Combined

S. borellii -31 Dec) at The Lakes Country Club (FS) and Sea Trail Plantati ) (%) Females No. Males (%) No. Total (%) Total

S. vicinus Fall mole cricket flight trap data by site, year, species, and sex from two sites in North Carolina for 1995 and 1996. Data cricket flight trap data by site, year, species, and sex from Fall mole FS 1995FS 1996 118 (95.2) 13 (100.0) 6 (4.8) 0 (0.0) 124 (68.9) 13 (9.3) 48 (85.7) 101 (79.5) 8 (14.3) 26 (20.5) 56 (31.1) 127 (90.7) 180 140 ST 1995ST 1996 0 (0.0) 6 (100.0) 0 (0.0) 0 (0.0) 6 (4.9) 0 (0.0) 92 (79.3) 65 (77.4) 24 (20.7) 19 (22.6) 84 (100.0) 116 (95.1) 84 122 Site Year (%) No. Females Males (%) No. Total (% Table 3. are numbers (%) captured in standard paired calling traps (1 Sept are numbers (ST), Brunswick Co., NC. 35 Scaptericsus Calendar date estimate for flight Calendar date estimate for hatch at two sites in North Carolina. (FS = The Lakes CC, ST Sea Trail Plantation). S. borellii Calendar date estimates for mean, 25, 50 and 75% cumulative spring flight and hatch for 25, 50 and 75% cumulative for mean, Calendar date estimates and FSFS 1995FS 1996 6 MayFS 1997 6 April May 6 1998 9 April May 5 15 May 7 April April 9 14 May 12 May 26 April April 9 May 3 13 May 14 June 16 AprilFS 26 April 19 June ----FS 12 June 1995 25 JuneFS 19 June 1996 ---- 2 May 20 June ----FS 27 June 1997 20 May May 6 19 June 1998 ---- June 9 3 June 25 May ---- 1 June 26 June 25 June 17 May ---- 18 June 7 June 8 July ---- 26 June 6 July 15 June ---- 4 July 5 July 14 June 9 July 12 July 10 July 15 July 15 July ---- 4 July 10 July 22 July ---- 15 July ------STST 1995 18 AprilST 1996 20 April 18 April 1997 12 May 29 March May 2 31 March 29 April 22 April May 6 10 April 20 June 28 April 25 June 18 JuneST 15 June 29 June 26 JuneST 1995 22 June 25 June 3 July 1996 14 May 28 June 25 June 23 May 22 June June 1 29 May 11 June 30 June 1 June 9 June ------Site Year 25% 50% 75% Mean 25% 50% 75% Mean Scapteriscus vicinus Scapteriscus borellii vicinus Table 4.

36 Figure 1. Cumulative spring flight counts of two species of mole crickets at two sites in

North Carolina, 1995 - 1998 (FS = The Lakes C. C., ST = Sea Trail Plantation).

37

29-Jul 29-Jul

14-Jul 14-Jul

29-Jun 29-Jun

14-Jun

14-Jun

30-May

30-May

15-May

15-May

30-Apr

30-Apr

15-Apr

15-Apr

31-Mar

31-Mar

16-Mar 16-Mar

T-ST95 T-ST96 T-ST97

S-ST95 S-ST96 S-ST97

1-Mar 1-Mar 0 0 50 400 350 300 250 200 150 100 500

2000 1500 1000

29-Jul 29-Jul

14-Jul 14-Jul

29-Jun 29-Jun

14-Jun 14-Jun

30-May

30-May

15-May

15-May

30-Apr

30-Apr

15-Apr

15-Apr

31-Mar

31-Mar

16-Mar

16-Mar S-FS95 S-FS96 S-FS97 S-FS98

T-FS95 T-FS96 T-FS97 T-FS98

1-Mar 1-Mar 0 0 50

400 350 300 250 200 150 100 500

2000 1500 1000

Cum. No. of S. vicinus S. of No. Cum. borellii S. of No. Cum.

38 Figure 2. Cumulative fall flight counts of two species of mole crickets at two sites in North

Carolina, 1995 - 1996 (FS = The Lakes C. C., ST = Sea Trail Plantation).

39 140 T-FS95 120 T-FS96 T-ST95 100 T-ST96

80

60

40 Cum. No. S. vicinus

20

0 1-Oct 1-Sep 16-Oct 31-Oct 15-Nov 30-Nov 15-Dec 30-Dec 16-Sep

140 S-FS95 120 S-FS96 S-ST95 100 S-ST96

80

60

40 Cum. No. of S. borellii 20

0 1-Oct 1-Sep 16-Oct 31-Oct 15-Nov 30-Nov 16-Sep 15-Dec 30-Dec

40 Chapter II

The Effect of Soil Moisture on Ovipositional Behavior in the Southern Mole Cricket

Scapteriscus borellii Giglio-Tos (Orthoptera: Gryllotalpidae)

Peter T. Hertl, Rick L. Brandenburg, and Mary E. Barbercheck

______

ABSTRACT The relationship between soil moisture and oviposition in an edaphic insect pest, the southern mole cricket, Scapteriscus borellii Giglio-Tos, was studied in a series of greenhouse experiments. Adults were captured in acoustic calling traps and associated pitfall traps during spring flights in southeastern North Carolina in 1996, 1997 and 1998.

Female mole crickets were individually confined in chambers containing 2, 4, 7, 10 and

12% soil moisture. Oviposition and mortality were monitored daily. A significant linear relationship between oviposition and soil moisture was indicated by an increase in the number of crickets ovipositing in response to higher soil moisture levels. Additionally, a delay in oviposition was observed as a response to low soil moisture. There were no significant differences in the number of eggs per ovipositing female, indicating that when oviposition does take place, the individual response of the female is to lay a similar number of eggs regardless of moisture levels. The ovipositional response to a rapid increase in soil moisture was also examined. The rapid increase in moisture resulted in a significantly greater percentage of females ovipositing, as seen in the previous experiments.

41 KEY WORDS Scapteriscus borellii, mole crickets, behavior, oviposition, soil moisture.

______

Mole crickets in the genus Scapteriscus are some of the most important insect pests of warm season turfgrasses in the southeastern United States. The crickets have a one year life cycle throughout most of their range and most of that time is spent in subterranean burrows.

Nocturnal mating and dispersal flights occur in the spring and fall and the eggs are laid in multiple clutches which develop within sealed chambers 2.5 to 30.5 cm below the soil surface (Hayslip 1943, Forrest 1986). Both the nymphs and adults cause serious damage to pastures, golf courses, home lawns, and other turfgrass sites, and successful control usually depends upon accurately timing the application of insecticides to coincide with the occurrence of small instars following peak egg-hatch in early summer (Brandenburg et al.

1997). However, field monitoring of mole cricket populations in North Carolina has shown that there can be significant fluctuations in the date when the first nymphs appear each year, and similar variation is seen in the date of peak hatch (P.T.H., unpublished data). These observed differences in the field are not well-explained by factors such as accumulated soil degree-days and are hypothesized to be due, in part, to differences in soil moisture levels during the period of oviposition.

Soil moisture could affect the timing of hatch through a direct effect on the eggs or by modifying the ovipositional behavior of the gravid females. Mole cricket eggs absorb water during development and increase in size as much as 25% prior to hatch (Hayslip 1943). If female mole crickets oviposit in the spring regardless of the soil moisture levels present at the time, then low soil moisture conditions could have a direct effect on the eggs, slowing or

42 halting their development until the necessary moisture becomes available. Another possibility is that the females may not oviposit until adequate soil moisture is available.

Previous investigations involving oviposition in Scapteriscus spp. include the natural history studies of Worsham and Reed (1912), Van Zwaluwenburg (1918) and Hayslip (1943), a study of fall mating and fertilization in S. borellii Giglio-Tos (Walker and Nation 1982),

Forrest’s (1986) work on oviposition and maternal investment in field enclosures, and a laboratory study examining egg production and development (Braman 1993). Only one previous study, an unpublished laboratory experiment conducted by Forrest (1983a), has attempted to relate egg survival and hatch with soil moisture for S. borellii. Although the initial moisture condition of the eggs was not reported, the results of this experiment indicate that soil moisture (between 2-12% of dry soil weight) did not influence hatching or development rate of the eggs. This suggests that soil moisture may affect the ovipositional behavior of the adults rather than the development of the eggs themselves. Unfortunately, there are at least two factors that complicate the study of these relationships in the field.

The first of these is subterranean nature of the mole cricket and the difficulty in making observations on oviposition and hatch. Equally important are the difficulties encountered in accurately establishing, maintaining, and manipulating soil moisture levels in field experiments. Because it is unlikely that these relationships can be adequately quantified through field studies, three controlled greenhouse oviposition experiments were conducted in 1996, 1997 and 1998. The objective of these experiments was to elucidate the relationship between soil moisture and ovipositional behavior of S. borellii.

43 Materials and Methods

All mole crickets used in the experiments were collected at the Fox Squirrel Country

Club in Brunswick County, NC (34E 2' N, 78E3' W). Male and female southern mole crickets were collected from sound-baited calling traps and linear pitfall traps set in the immediate vicinity of the calling traps. The calling traps use electronically synthesized species-specific songs to attract flying adults (Walker 1982). These traps capture mainly females. Most of the males were collected from the linear pitfall traps. Mole crickets in both types of traps were captured into approximately 10 liters of moist sand prior to their removal from the traps. Collections in 1996 (Experiment I) were made during the spring flights, which occurred in the latter part of May, just prior to oviposition in the field. Mole crickets used in the four blocks of the 1997 experiment (Experiment II) were captured in four separate collections made at approximately 1- wk intervals (beginning 14 May, and ending 4 June). The two collections for the 1998 experiment (Experiment III) were made on

9 and 19 June. In all cases the mole crickets were captured no more than 1 wk prior to their removal from the traps. The effect of mating on oviposition in mole crickets is unknown, however, research conducted in Florida indicates that most of the flying females of this species have already mated prior to the spring flights (Walker and Nation 1982). To further ensure that mating had taken place, both sexes were confined together in large coolers full of moist sand for a few days prior to the initiation of each experiment. All soil used in the experiments was Kureb fine sand obtained from the same location where the insects were collected. To enable the comparison of the soil moisture levels used in these experiments with those in future studies, a standard matric potential curve for this soil was developed

44 using the standard pressure plate extractor technique following the methods described by

Kramer (1969). Voucher specimens of S. borellii were placed in the North Carolina State

University insect collection.

Experiment I. In 1996 a preliminary experiment was conducted in a greenhouse with the treatments set up in a completely random design. Soil moisture regimens were established prior to the start of the experiment by combining a weighed portion of distilled water with a weighed amount of oven-dried sand. This procedure was used to produce soil mixtures with 2, 7, and 12% moisture based on the dry weight of the soil. Each treatment consisted of 20 PVC oviposition chambers 7.7-cm diameter, 15-cm long similar to those used by Braman (1993). Chambers were filled with 12.7 cm of one of the soil mixtures and one randomly selected female mole cricket was confined in each chamber along with a portion of Mazuri Hi-Ca Cricket Diet (Purina Mills, Inc.) treated with 0.4% potassium sorbate to inhibit fungal growth. The oviposition chambers were capped on each end with a clear plastic petri dish lid placed over a layer of plastic wrap, then secured with a rubber band to contain the insects, the soil, and the soil moisture. Each chamber was then weighed and soil moisture levels were maintained throughout the experiment by weighing the tubes weekly and adding an appropriate amount of water to maintain the starting moisture level.

Mole crickets were fed weekly by adding pre-weighed portions of food. A mock chamber with a thermometer was used to monitor chamber temperature.

Oviposition was monitored by opening and inspecting the bottom of each enclosure for the presence of eggs. The number of eggs laid in each chamber was recorded daily. The process of checking for the eggs usually disrupts the egg chamber formed by the female,

45 making the accuracy of later observations uncertain. S. borellii is primarily a carnivorous species (Matheny 1981) and cannibalism of the eggs frequently occurs under laboratory conditions (Van Zwaluwenburg 1918). Therefore, after eggs were initially noted and counted in a chamber, further observations of oviposition were not included in the final counts. The ovipositional responses to the soil moisture treatments in all experiments were quantified by four measures. These were 1) the mean number of eggs laid per female (total eggs laid divided by the 20 females in each treatment), 2) the mean number of eggs laid by each female that laid eggs (total eggs laid divided by the number of females ovipositing), 3) the percentage of females that oviposited in each treatment (number of ovipositing females divided by number of females in the treatment), and 4) the mean day that eggs were laid in each treatment.

Obvious adult mortality (death on the soil surface or bottom of chamber) was recorded daily, and inactive chambers were occasionally checked to determine if the insect was dead.

Experiment I was terminated on d 31 due to suspected high mortality in the 2% moisture treatment. The soil from the chambers in this treatment was sifted using a 20 mesh sieve and both the sieved material and the dead mole crickets were returned to the laboratory and examined for the presence of eggs using a dissecting scope. The presence or absence of eggs in the sand was recorded and data regarding percentage of females ovipositing, eggs per female, eggs per ovipositing female, mortality, and mean day of oviposition were summarized to d 31. While this experiment proved invaluable in designing the later experiments, high mortality in one treatment, and the unblocked nature of the experimental design precluded statistical analysis by ANOVA or mean separation procedures.

46 Experiment II. This experiment was set up in a randomized complete block design with

20 chambers per treatment in four blocks. Soil for each moisture treatment (4, 7, and 10% based on dry soil weight) was prepared as in Experiment I, and each block of the experiment began approximately 1 wk after the start of the previous block. Three of these blocks were performed in the same greenhouse as Experiment I. The fourth block was performed in an indoor temperature controlled facility under artificial lighting. Light cycles were set to approximate the light/dark cycle existing outdoors at the time of the experiment.

Observations on oviposition were made daily and chambers were maintained as described in

Experiment I. Obvious adult mortality was recorded daily and inactive chambers were checked weekly by sieving the soil and emptying the chambers to determine if the cricket was dead. This procedure was generally performed at the same time as the weekly feeding and moisture level adjustment. Mortality reported at dates prior to the end of the experiments must be considered as estimated mortality due to the difficulty in making these observations. Mortality will never be lower than reported, but could be higher in cases where the death was not noted until later in the experiment. Experiment II was run for 32 d and cumulative data regarding mortality, percentage of females ovipositing, eggs per female, and eggs per ovipositing female were summarized at five dates (d 7, 14, 21, 28 and 32).

Data regarding mean day of oviposition were only summarized at d 32. Data were analyzed using either an arcsine (percentage data) or square root transformation (count data) prior to

ANOVA and LSD procedures. Significance for mean separation was tested using either

LSD or LSD performed on Least Square Means (LSmeans) (for unbalanced data), as appropriate (SAS Institute 1992). All means are presented as untransformed means ± SE.

47 Experiment III. This experiment was performed in order to test the hypothesis that females are stimulated to oviposit by a rapid increase in soil moisture. The experiment was set up in the greenhouse as a randomized complete block design with two blocks and 35 chambers in each of two treatments. Soil and water were combined as previously described to produce soil with 7% moisture based on dry soil weight. Chambers were assembled as before, but all contained 7% soil moisture at the start of the experiment. Female mole crickets were individually placed into the chambers, fed, and monitored as before, for the first 6 d. After the chambers were checked for eggs on d 6, each of the chambers in the flush treatment were flushed by adding 40 ml of distilled water, increasing the soil moisture level to approximately 13%. Food was added and then the chambers were resealed.

Chambers in the control treatment were fed and moisture was adjusted to maintain the starting 7% moisture level, but no additional water was added. All chambers were monitored daily for 10 d after the flush. Data regarding the percentage of females ovipositing, eggs per female, eggs per ovipositing female, mortality, and mean day of oviposition for each treatment were analyzed using ANOVA and significance for mean separation was tested using either an F or Chi-square test, as appropriate (SAS Institute

1992).

Results

Moisture and Temperature. The soil moisture in the 2% moisture treatment used in

Experiment I was not sufficient to maintain tunnel, egg chamber, or soil column integrity.

The loose soil made observations on oviposition and adult survival difficult and undoubtedly affected the observations for this treatment. Additionally, there was extremely

48 high mortality in the 2% moisture treatment. Higher levels of moisture were used in all subsequent experiments and no further problems of this nature were encountered.

Soil temperatures encountered in the greenhouse were similar to those observed at a depth of 10 cm in the field during the time of oviposition (P.T.H., unpublished data), but the mole crickets confined in the oviposition chambers were not able to escape to cooler depths as they can in the field. In Experiment I temperatures ranged between 20 and 35.6EC with most recorded temperatures in the mid to high twenties. Temperatures in blocks I-III of

Experiment II ranged from 15.6 to 33.3EC with most recorded temperatures in the mid to high twenties. Block IV of that experiment was performed in a more controlled environment in an attempt to avoid high temperatures. Temperatures observed ranged from

21.1 to 24.4EC. Chamber temperature in Experiment III ranged between 23.3 and 32.2EC with most recorded temperatures in the upper twenties.

Mortality. Determination of mortality effects was not an objective of these experiments, and observations were made primarily to ensure that mortality had no effect on the analysis of egg-laying among the treatments. Some mortality was expected because the experiments were performed at the time of year when the adult population is in decline. Although mortality is the natural conclusion to the life cycle after the last clutch of eggs are laid, it should have been equally distributed among the treatments. However, major differences in mortality were noted among the treatments in Experiment I. Although there were difficulties in making accurate observations in the 2% treatment at least one mole cricket was dead by d 8, two more were found dead on d 13, and at least 20% (5 of 20) of the mole crickets were dead by d 23. Mortality in the 2% moisture treatment was 100% by d 31, but

49 the decomposed condition of the cadavers strongly suggests that many of these insects died much earlier. It is likely that the combination of low moisture and high temperatures contributed substantially to the observed mortality. The high mortality observed in the 2% moisture treatment was sharply contrasted by the high survival in the 7 and 12% moisture treatments in this experiment. No mortality was observed in the 12% moisture treatment and only one mole cricket (5%) in the 7% moisture treatment died before d 31.

No significant differences in mortality were observed among the blocks or treatments in

Experiments II (Table 1) at any date. This indicates that even the moisture present in the

4% treatment was sufficient to allow the survival of the adults. Mortality at d 32 ranged between 20 and 53% with an overall mean value of 32.0 ±3.3%. There were significant differences in mortality between the blocks (F-test, P=0.0001) but not between the treatments (F-test, P = 0.65) in Experiment III (Table 2).

Experiment I. Oviposition. The average number of eggs per female , the number of eggs per ovipositing female, and the percentage of females ovipositing were all highest in the 12% moisture treatment and lowest in the 2% treatment. It was not possible to determine when the two eggs in the 2% moisture treatment were laid, but oviposition began

2.5 d earlier in the 12% treatment than in the 7% treatment. Large differences were seen in the average number of eggs per female in the 7 and 12% moisture treatments (5.5 and 12.0 eggs per female, respectively). The numbers of eggs produced per ovipositing female show similar trends. Only one egg was laid by each of the two females ovipositing in the 2% moisture treatment while 13.8 and 17.1 eggs were laid per ovipositing female in the 7 and

12% treatments, respectively. Additionally, 40.0% and 70.0% of the females laid eggs in

50 the 7 and 12% moisture treatments, respectively. Only 10.0% of the females oviposited in the 2% moisture treatment. These data suggest that increasing soil moisture has a positive effect on the number of eggs laid and the proportion of females ovipositing.

Experiment II. Percentage of Females Ovipositing. As there were no significant differences in mortality among the treatment or blocks of Experiment II, oviposition data for all blocks and treatments could be compared and were included in the analysis. Data regarding the percentage of females ovipositing in each treatment are presented graphically in Fig. 1. Most egg-laying activity in Experiment II took place after d 14, and statistical analysis was not performed prior to d 21. At d 21 there were significantly more eggs laid in the 10% moisture treatment than in either the 4 or 7% moisture treatment (LSD, P = 0.05); however, no differences were detected in the percentage of females laying eggs between the

4 and 7% treatments (LSD, P = 0.05). The percentage of mole crickets ovipositing in each treatment at 28 and 32 d are presented in Table 1. At 28 d, 29.2 ± 0.07% of all mole crickets in the experiment had laid eggs. There were significant differences in the mean number of females ovipositing among all treatments, with the greatest proportion of females (51.3%) laying eggs in the 10% moisture treatment, the smallest proportion of females (6.3%) laying eggs in the 4% treatment, and an intermediate number of females (30.0%) ovipositing in the

7% treatment. At d 32 an average of 36.4 ± 0.07% of all mole crickets in the experiment had laid eggs. The mean values for each treatment reflect the same general trends observed at previous dates. While the results of the LSD means separation test indicate no significant differences in the proportion of females laying eggs in the 7 and 10% treatments at this date, there were significantly fewer females ovipositing in the 4% treatment than in either of the

51 other moisture treatments (LSD, P =0.05).

The relationship between soil moisture and the percentage of females ovipositing in

Experiment II (Fig. 2) is linear as described by the equation y = -8.70 + 6.43x , r2 = 0.9736

(y = a + bx, where y = percentage of females ovipositing, x = percentage soil moisture). The linear relationship is strong, as evidenced by the high r2 value. Although the data from

Experiment I were not used in the calculation of the regression line, the observed values for percentage oviposition in the 7 and 12% moisture treatment in that experiment (40 and 70%) agrees with the predicted value of 36.31 and 68.50%. The predicted value of 4.17% oviposition for the 2% treatment was not as close to the 10% observed value, but this figure only represents two females laying a single egg each.

Mean Number of Eggs per Female. Results of data analysis for the mean number of eggs per female in Experiment II (Table 1) closely follow the results obtained for the percentage of females ovipositing. There were no significant differences observed among the treatments at d 14. At d 21 there were significantly more eggs laid per female in the 10% moisture treatment than in the 4% treatment. Mean number of eggs per female in the 7% treatment was not statistically different from that observed in either of the other treatments.

At d 28 all treatments were significantly different from each other with regard to the number of eggs per female (LSD, P = 0.05). The response was an increasing number of eggs per female as soil moisture increased. Although there were only 1.2 eggs laid per female in the

4% moisture treatment, there were 5.8 and 11.1 eggs laid per female in the 7 and 10% treatments, respectively. At d 32 there was an overall average of 7.25 ± 1.43 eggs per female with a similar relationship between the number of eggs per female and soil moisture

52 levels obvious among the treatment means. At d 32 the number of eggs in the 7 and 10% treatments were not significantly different; however, significantly fewer eggs were laid in the 4% moisture treatment than either the 7 or 10% treatments.

Mean Number of Eggs per Ovipositing Female. Means separation for the number of eggs per ovipositing female in Experiment II is only reported for d 28 and 32 due to the lack of egg-laying in some blocks or treatments at prior dates. Results for d 32 indicate that there was an overall average of 19.3 ± 1.0 eggs laid by those females laying eggs. There were no significant differences observed among either the blocks or treatments at either d 28 or 32

(LSD, P = 0.05) indicating that the moisture treatments influence whether or not a female laid eggs, rather than the number of eggs the female laid.

Timing of Oviposition. Mean day of oviposition was examined to determine if the timing of oviposition was affected by soil moisture. Analysis at d 32 shows the overall mean oviposition date for Experiment II was d 24.6 ± 1.1. There were significant differences observed among both blocks and treatments regarding the timing of oviposition (Table 1).

Eggs were laid an average of 3.6 d earlier in the 10% moisture treatment than in the 4% treatment.

Experiment III. This experiment was performed to determine if gravid female S. borellii would be stimulated to oviposit by a rapid increase in soil moisture. Significantly more females oviposited in the flush treatment, resulting in significantly more eggs being laid (Table 2 and Fig.3), however, there was no significant difference in the mean day of oviposition between the two treatments. No differences were seen in mean number of eggs laid per ovipositing female. These results indicate that while increased soil moisture levels

53 increased the percentage of crickets ovipositing (as observed in the previous experiments), the timing of oviposition was not significantly affected.

Discussion

Soil moisture has long been recognized as an important environmental factor that influences the behavior of mole crickets. Hayslip (1943), Ulagaraj (1975) and Walker

(1982) have all cited rainfall as stimulating adult flights, an activity that often immediately precedes oviposition. Ulagaraj (1976) found that moisture influenced the production of calling songs in both S. borellii and S. vicinus Scudder in a field irrigation experiment where males of both species were found singing in the irrigated plots, but not in the unirrigated plots. High soil moisture significantly increases the power output (sound volume measured in decibels) of the male call (Forrest 1991), and Forrest (1983b) noted that females may use the song intensities of the calling males to locate moist soil for oviposition. Flying female mole crickets selectively respond to louder males and the loudest S. borellii males attracted

6.5 times more conspecifics than quieter males (Forrest 1983b). Soil moisture also influences the depth at which the females form their egg chamber. Scapteriscus egg cells formed during dry soil conditions were found at two to three times the depth as those produced during moist soil conditions ( Worsham and Reed 1912, Van Zwaluwenburg 1918,

Hayslip 1943). These studies suggest that moisture is an important factor influencing the selection of mole cricket oviposition sites.

Our results suggest that gravid females prefer to oviposit in areas with adequate moisture and avoid or delay oviposition where dry soil conditions exist. In Experiment I, over twice the number of eggs were laid in the 12% treatment as in the 7% moisture treatment. In

54 Experiment II, significantly more crickets oviposited in the higher moisture treatments (7 and 10%) than in the 4% treatment, resulting in significantly more eggs being laid in these treatments. Similarly, the results of Experiment III show a significant increase in the percentage of females ovipositing in response to a rapid increase in soil moisture. The results of all three experiments show that soil moisture levels influence whether or not a female will lay eggs, rather than the number of eggs the female will lay.

The timing of oviposition was another important variable that might have been affected by soil moisture levels. In Experiment I, oviposition began 2.5 d earlier in the 12% treatment than in the 7% treatment. Only two eggs were found in the soil of the 2% moisture treatment, however, examination of the cadavers showed that these females were gravid, but did not oviposit. In Experiment II, the mole crickets laid eggs an average of 3.6 d earlier in the 10% moisture treatments than in the 4% treatment. No eggs were laid in the

4% moisture treatment of block IV, which was run in a temperature-controlled environment, but oviposition did occur in the other treatments of this block. The lack of oviposition here is attributed to the treatment effect of low soil moisture under less stressful temperature conditions. Although these data are meaningfully represented in the analyses of eggs per female and percentage of females ovipositing, they could not be included in the calculation of the average value for day of oviposition because oviposition did not occur. This resulted in the delay in oviposition due to low soil moisture being less obvious than it might have been if this factor was considered. The results from Experiment III did not show the differences in timing observed in the previous experiments, however, those experiments were of longer duration, and the delay in oviposition only became obvious at moisture levels

55 lower than the 7% treatment that served as a control in Experiment III.

The ovipositional data from our study were also used to determine if there is a correlation between ovipositional cycles and flight activity. In a study on the dispersal flights of S. borellii in Florida, Ngo and Beck (1982) found 9-d cycles in overall flight activity and a 10- d peak in recaptured females in a mark and release study. They attributed both peaks in flight activity to the 9-12 d oviposition cycles reported by Forrest (1981), a conclusion that assumes that the females are synchronous in oviposition (Forrest 1986). In a later study,

Forrest (1986) found no such synchrony and concluded that the peaks in flight activity were more likely the result of environmental factors making flight favorable or synchronizing maturation of the eggs, rather than ovipositional cycles. In our study we used females captured in flight traps. These females were often captured several days prior to collection, then held in the laboratory for 3 d prior to initiation of the experiment. Only 15 of the 440

(3.4%) mole crickets in our experiments oviposited prior to d 14, which means that a minimum of 17 d had elapsed since their capture. Therefore, we agree with Forrest’s conclusion that cycles in flight activity are more related to environmental factors than ovipositional cycles. Soil moisture is very likely one of these factors. Although our results demonstrate a moisture-mediated delay in oviposition, we do not discount the possibility of an additional moisture effect on the survival and development of mole cricket eggs to explain the population variation observed in the field. Van Zwaluwenburg (1918) reported that Scapteriscus spp. eggs cannot survive in a very dry location and do not hatch after dessication, which suggests that there is such an effect. There are no published studies relating soil moisture to mole cricket egg survival and development, however, Regniere et

56 al. (1981) studied egg survival and development of another soil-dwelling pest, the Japanese beetle larva, Popillia japonica Newman. They found that egg survival was not possible in sandy and loam soils below 3% soil moisture and presented a moisture tension graph similar to the matric potential graph (Fig. 4) for the soil used in our experiments. Matric potential is the primary factor determining the availability of soil water to plants and animals (Hillel

1980, Barbercheck 1992) and the graph in Fig. 4 clearly shows that water is biologically unavailable at 2% in our soil. Although Krysan (1976) and many others have documented the negative effects of low relative humidity (RH) on the eggs and larvae of soil insects, data presented by Brady (1974) indicate that the RH of the soil atmosphere in the sealed egg chamber of the mole cricket would remain near 100% up to -100 KPa of water potential.

This is approximately equivalent to 3% soil moisture for the soil used in our experiments.

Brady’s (1974) data also explain why Forrest (1983a) found no effect of moisture on mole cricket egg survival. The eggs were kept in sealed jars, effectively mimicking the conditions of the sealed egg chamber. Even at the lowest soil moisture treatment in

Forrest’s experiment (2%), the RH of the atmosphere in the jars would have been in excess of 98%. Although our observations indicate that mole cricket eggs desicate very rapidly when exposed to the air, it is likely that they are not adversely affected by RH in this range once they have absorbed sufficient moisture for development.

The increase in the volume of the eggs after oviposition (Hayslip 1943) indicates that they must absorb water. Moisture absorption by mole cricket eggs has not been studied, however, Hinton (1981) reported that insect eggs only absorb water in the liquid phase, and

Mihm et al. (1974) found that contact moisture was necessary for egg hatch in Diabrotica

57 virgifera LeConte. Relatively rapid uptake of moisture has been documented for the eggs of other soil-dwelling insects, and Krysan (1976) reported that most of the uptake of water by the eggs of Diabrotica undecimpunctata howardi Barber occurred between 24 and 72 hr after oviposition. If moisture absorption for mole cricket eggs follows a similar pattern, the selection of a favorable moisture environment at oviposition would be critical to the survival of the eggs.

Host-habitat and host finding of soil-dwelling species begin with ovipositional choices made by the mobile females (Villani and Wright 1990). The advantages of selecting a favorable soil moisture environment for oviposition are obvious. Studies have demonstrated that soil moisture levels can have a significant effect on the survival and development of both the eggs and larvae of edaphic insects (Regniere et al. 1981, Marrone and Stinner

1983a), and adult ovipositional preference for moist soil has also been demonstrated

(Marrone and Stinner 1983b). There is no parental care reported for the genus Scapteriscus as has been documented for other gryllotalpids, which suggests that the survival of

Scapteriscus eggs is dependent upon the conditions existing in the habitat selected by the female at the time of oviposition. The importance of contact moisture rather than RH, and the likelihood of a critical period of water absorption makes a strong case for the evolutionary importance of initial conditions and the female behavior seen in our experiments. Both the preference exhibited for higher soil moisture levels at oviposition, and delaying oviposition until adequate moisture is available, reduce the possibility that dry soil conditions will impact egg survival or development.

58 Acknowledgments

The authors would like to extend their sincere thanks to superintendents Rick Vigland and

Chuck Baldwin of the Fox Squirrel Country Club, and student research assistants Ian

Winborne and Ed Karoly. Special thanks are also extended to Dr. Cavell Brownie (Dept. of

Statistics, NC State University) for indispensable assistance with the statistical analysis of results. This work was funded in part by a grant from the United States Golf Association

Greens Section Research.

References Cited

Barbercheck, M. E. 1992. Effects of soil physical factors on biological control agents of soil insect pests. Fla. Entomol. 75: 539-548.

Brady, N. C. 1974. The nature and properties of soils. MacMillan, New York.

Brandenburg, R. L., P. T. Hertl and M. G. Villani. 1997. Integration and adoption of a mole cricket management program in North Carolina, USA. Intl. Turf. Soc. Res. J. 8: 973-

979.

Braman, S. K. 1993. Progeny production, number of instars, and duration of development of tawny and southern mole crickets (Orthoptera: Gryllotalpidae). J. Entomol. Sci. 28: 327-

330.

59 Forrest, T. G. 1981. Acoustic behavior, phonotaxis and mate choice in two species of mole crickets (Gryllotalpidae: Scapteriscus). M. S. thesis, University of Florida,

Gainesville.

Forrest, T. G. 1983a. Survival of eggs as a function of soil moisture. Ann. Rep. No.5,

Mole Cricket Res. 82-83., Dept. of Ent. and Nem., University of Florida, Gainesville.

Forrest, T. G. 1983b. Calling songs and mate choice in mole crickets. pp. 185-204. In D.

T. Gwynne and G. K. Morris [eds.], Orthopteran mating systems: sexual competition in a diverse group of insects. Westview Press, Boulder, Colorado.

Forrest, T. G. 1986. Oviposition and maternal investment in mole crickets (Orthoptera:

Gryllotalpidae): effects of season, size, and senescence. Ann. Entomol. Soc. Am. 79: 918-

924.

Forrest, T. G. 1991. Power output and efficiency of sound production by crickets. Behav.

Ecol. 2(4): 327-338.

Hayslip, N. C. 1943. Notes on biological studies of mole crickets at Plant City, Florida.

Fla. Entomol. 26: 33-46.

Hillel, D. 1980. Fundamentals of soil physics. Academic Press, New York.

60 Hinton, H. E. 1981. Biology of insect eggs. Pergamon Press, Oxford.

Kramer, P. J. 1969. Plant and soil water relationships: a modern synthesis. McGraw-Hill,

New York.

Krysan, J. L. 1976. Moisture relationships of the egg of the southern corn rootworm,

Diabrotica undecimpunctata howardi (Coleoptera: Chrysomelidae). Entomol. Exp. Appl.

20: 154-162.

Marrone, P. G., and R. E. Stinner. 1983a. Effects of soil physical factors on egg survival of the bean leaf beetle, Cerotoma trifurcata (Foster) (Coleoptera: Chrysomelidae). Environ.

Entomol. 12: 673-679.

Marrone, P. G., and R. E. Stinner. 1983b. Effects of soil moisture and texture on oviposition preference of the bean leaf beetle, Cerotoma trifurcata (Foster) (Coleoptera:

Coccinellidae). Environ. Entomol. 12: 426-428.

Matheny, E. L. 1981. Contrasting feeding habits of pest mole cricket species. J. Econ.

Entomol. 74: 444-445.

Mihm, J. A., H. C. Chiang, and M. B. Windels. 1974. Moisture relationships of developing corn rootworm eggs, pp. 141-143. In Proceedings, Annual Meeting of the North

61 Central Branch, Entomological Society of America, Lanham, MD.

Ngo D., and H. W. Beck. 1982. Mark-release of sound-attracted mole crickets: flight behavior and implications for control. Fla. Entomol. 65: 531-538.

Regniere, J., R. L. Rabb, and R. E. Stinner. 1981. Popillia japonica: Effects of soil moisture and texture on survival and development of eggs and first instar grubs. Environ.

Entomol. 10: 654-660.

SAS Institute. 1992. SAS/STAT user’s guide, version 6.12. SAS Institute, Cary NC.

Ulagaraj, S. M. 1975. Mole crickets: ecology, behavior, and dispersal flight (Orthoptera:

Gryllotalpidae: Scapteriscus). Environ. Entomol. 4: 265-273.

Ulagaraj, S. M. 1976. Sound production in mole crickets (Orthoptera: Gryllotalpidae:

Scapteriscus). Ann. Entomol. Soc. Am. 69: 299-306.

Van Zwaluwenburg, R. G. 1918. The changa or West Indian mole cricket. Porto Rico

Agric. Exp. Stn. Bull. 23: 1-28.

Villani, M. G. and R. J. Wright. 1990. Environmental influences on soil macroarthropod behavior in agricultural systems. Annu. Rev. Entomol. 35: 249-69.

62 Walker, T. J. 1982. Sound traps for sampling mole cricket flights (Orthoptera:

Gryllotalpidae: Scapteriscus). Fla. Entomol. 65: 105-109.

Walker, T. J., and J. L. Nation. 1982. Sperm storage in mole crickets: fall matings fertilize spring eggs in Scapteriscus acletus. Fla. Entomol. 65: 283-285.

Worsham, E. L., and W. V. Reed. 1912. The mole cricket. Georgia Exp. Stn. Bull. 101:

251-263.

63 Table 1. Oviposition by S. borellii confined in chambers containing three different soil moisture treatments in Experiment II, summarized at day 28 and 32

Mean ± SE

Soil

moisture % females No. eggs No. eggs per % Day of

treatment ovipositing1 per female2 ovipositing mortality1 oviposition2

female3 Day 28

10 % 51.3 ± 9.7a 11.1 ± 1.9a 21.8 ± 0.7a 21.3 ± 4.7a --- 7 % 30.0 ± 9.4b 5.8 ± 1.9b 18.1 ± 2.4a 17.5 ± 4.8a --- 4 % 6.3 ± 3.8c 1.2 ± 0.7c 19.5 ± 4.5a 27.0 ± 8.6a --- Day 32

10 % 53.8 ± 5.4a 11.5 ± 2.0a 21.6 ± 0.6a 33.8 ± 6.3a 22.6 ± 1.6a 7 % 40.0 ± 8.7a 7.8 ± 1.9a 19.4 ± 2.1a 28.8 ± 5.2a 25.4 ± 2.0ab 4 % 15.2 ± 5.4b 2.4 ± 0.8b 16.1 ± 1.4a 33.5 ± 7.1a 26.2 ± 2.8b Data are means ± SE of four replicates with 20 females per treatment in each replicate.

Means in the same column followed by the same letter are not significantly different.

1 (LSD, P = 0.05): Arcsine transformation of percentage data prior to ANOVA.

2 LSD comparisons performed on LSMeans due to unbalanced data, (P = 0.05).

3 (LSD, P = 0.05).

64 Table 2. Oviposition response of S. borellii subjected to a rapid increase in soil moisture in Experiment III, summarized at 10 day after the flush

Mean ± SE

Soil % females No. eggs No. eggs per % Day of moisture ovipositing1 per female2 ovipositing mortality2 oviposition2 treatment female2

Flush 50.8 ± 6.2a 7.7 ± 1.1a 15.1 ± 1.0a 26.9 ± 5.5a 7.7 ± 0.3a Control 17.9 ± 4.7b 2.9 ± 0.8b 16.3 ± 1.7a 29.9 ± 5.6a 8.2 ± 0.4a

Data are means ± SE of two replicates with 35 females per treatment in each replicate.

Means in the same column followed by the same letter are not significantly different.

1 Chi-square Test, (P = 0.012).

2 F-Test, (P = 0.05).

65 60

50 4%

40 7% 10% 30

20

10

% FEMALES OVIPOSITING 0 1 4 7 10 13 16 19 22 25 28 31 DAYS

Figure 1. Comparison of the cumulative ovipositional response of

female S. borellii among the three soil moisture treatments in

Experiment II (4 blocks, n = 20).

66 60

50

40

30

20

10 % FEMALES OVIPOSITING 0 024681012 % SOIL MOISTURE

Figure 2. Relationship between soil moisture and mean percentage

of female S. borellii ovipositing in the three moisture treatments

in Experiment II.

67 8.0

7.0

Control 6.0 Flush 5.0

4.0

3.0

2.0

1.0 MEAN EGGS PER FEMALE

0.0 12345678910 DAYS AFTER FLUSH

Figure 3. Comparison of the cumulative ovipositional response of female S. borellii in the two treatments of Experiment III after a rapid increase in soil moisture in the flush treatment (2 blocks, n = 35).

68 1600

1400

1200

1000

800

600

400

200

MATRIC POTENTIAL (- KPa) 0 0 2 4 6 8 1012141618202224 % SOIL MOISTURE

Figure 4. Relationship between soil moisture and matric potential

(-KPa) for the soil used in the S. borellii oviposition experiments.

69 Chapter III

The Development of Scapteriscus vicinus Scudder and S. borellii Giglio-Tos

(Orthoptera: Gryllotalpidae) in southeastern North Carolina.

Peter T. Hertl, Rick L. Brandenburg, R. E. Stinner and Cavell Brownie

______

ABSTRACT The development of the tawny mole cricket (Scapteriscus vicinus Scudder) and southern mole cricket (S. borellii Giglio-Tos) was quantified in southeastern North

Carolina from field-collected samples. Nymphs were sampled weekly using the soapy water flush technique during the summers of 1993–1997 at nine golf courses to compile a development data base from a total of 20 site-years. Pronotal length was used to assign the nymphs to size classes, and counts were summarized on a m2 basis. The two smallest size classes were equated to the first and second instar and counts were used separately to quantify the timing of 25, 50 and 75% peak and cumulative abundance. These data were further used to estimate the timing of oviposition and hatch. On-site soil degree-day accumulations and rainfall data from local weather stations were examined to determine their relationship to the timing of development. Soil degree-days were correlated with the timing of cumulative nymph abundance, however, calendar date quantified timing better than degree-day estimates. A statistically significant relationship between nymph abundance and both soil degree-days and rainfall was found. Differences in annual degree- day accumulations and a soil moisture-related delay in oviposition documented in previous

70 greenhouse experiments are believed responsible for differences in annual development observed in the samples. Management implications of the research are discussed.

KEY WORDS Scapteriscus vicinus, Scapteriscus borellii, mole crickets, development, degree-days, soil moisture, turfgrass.

______

Two species of introduced mole cricket pests, Scapteriscus vicinus Scudder and S. borellii Giglio-Tos (formerly S. acletus Rehn and Hebard), cause extensive tunneling and feeding damage to home lawns, golf courses, and other turfgrass throughout the southeastern

U.S. coastal states (Walker 1984, Braman 1993). S. vicinus is primarily a phytophagous species, feeding on the roots and stems of the grass. S. borellii is primarily a predaceous species, feeding mainly on other soil arthropods, but also feeding occasionally on the grass

(Matheny 1981). Both species have a similar one year life cycle in North Carolina, but there are some significant differences in the timing of flights, oviposition, egg-hatch and development. Flight activity begins in March and continues well into June and July (Hertl et al. 2003). Most oviposition takes place in May and June. Eggs are deposited in underground chambers formed by the female with 1–59 eggs per cluster (Hayslip 1943).

Females lay up to ten clutches and can produce more than 450 eggs during their lifetime

(Forrest 1986). The adults die soon after oviposition. The eggs hatch in June and July, and the young nymphs make their way into the upper root zone and begin to feed. Although all active stages are destructive to the turf, the damage usually does not become apparent until early August when the nymphs are larger (Short and Koehler 1979, Hertl et al. 2002). The two species differ significantly in the proportion of the population that overwinter as

71 nymphs. The majority of the S. vicinus population (75–85%) develop to the adult stage by

December, while approximately the same proportion of S. borellii overwinter as nymphs and become adults the following spring.

Management of these turfgrass pests requires a large investment in time, labor, equipment, and insecticides. Insecticides are currently the only management strategy that provides acceptable control, however, these treatments are often ineffective due to improper timing (Brandenburg and Williams 1993). The timing of insecticide applications is an important consideration because insecticides vary considerably in effective residual time, mode of action and insect stage targeted. Insecticides applied against nymphs are usually most successful when the population is primarily composed of small susceptible instars, and will only provide good control if the residual effect of the treatment lasts until the remaining eggs have hatched. However, some of the more popular, broad spectrum insecticides have a short half-life and are ineffective on the larger nymphs, making the timing of controls critical. This means there is only a narrow window of opportunity to implement control.

Development of a comprehensive IPM program requires a substantial amount of information about the development, distribution, density, behavior, and physiology of target pests. This is especially important in developing methods for predicting the occurrence of various life stages targeted for control. Prior to this study, little was known about mole cricket biology in North Carolina, including the information necessary to develop such a program. The objective of this research was to quantify mole cricket development in southeastern North Carolina and develop a phenological model of population development as an aid to making mole cricket management decisions.

72 Materials and Methods

Systematic mole cricket sampling began in June 1993 and was continued during the summer months each year until July 1997. Virtually all the samples considered here were taken on managed bermudagrass (Cynodon dactylon (L.) Pers.) golf course fairways, but some sampling also occurred on tees, driving ranges or in areas of the rough. In all cases the courses were managed following standard management procedures and no attempt was made to change mowing, fertilization, irrigation or insect management practices at the courses sampled. Unless otherwise noted, the sampling procedure was the same at all locations.

Sampling Sites and Years Sampled. The study sites were located in Brunswick, New

Hanover and Pender Counties which represent the southeastern coastal extreme of North

Carolina. Brierwood Golf Club (BW), Shallotte, NC (33E 57' N, 78E 24' W)- An 18 hole bermudagrass golf course located in mid Brunswick County. Soil types and level of infestation vary substantially among the fairways. This site was sampled weekly in 1996.

Cape Fear Country Club (CF), Wilmington, NC (34E 13' N, 77E 55' W)- A highly managed 18 hole bermudagrass golf course with a limited mole cricket population. This site was sampled for mole crickets at two week intervals in 1993.

Duck Haven Golf Club (DH), Wilmington, NC ( 34E 14' N, 77E 51' W)- An 18 hole public golf course located in northern Wilmington near LF. This is a low input bermudagrass course with heavier soils infested primarily by S. borellii. This site was sampled weekly in

1994 and 1995.

The Lakes Country Club (FS, formerly the Fox Squirrel Country Club), Boiling Spring

73 Lakes, NC (34E 02' N, 78E 03' W ) - An 18 hole golf course located approximately 28 km southeast of Wilmington NC. The course has had a serious mole cricket infestation since approximately 1987 with the damage becoming more severe each year. The sandy soil is classified as Kureb, Mandarin, and Leon fine sand with 0.5–0.7% humic matter and a pH of

5.2–5.9. The fairways are primarily Tifgreen 328 bermudagrass, but vary somewhat in percentage weed infestation, soil compaction, and management practices. This site was sampled for mole crickets weekly every summer from 1993–1997. Concurrent studies of mole cricket flight activity (1995–1998) and surface activity (1995–1997) were also conducted at this site.

Landfall (LF), Wilmington, NC (34E 14' N, 77E 49' W)- A complex of three 18 hole golf courses with bermudagrass fairways and centipedegrass roughs located in the northern portion of Wilmington. The course was sampled at varying intervals in 1993.

Olde Point (OP), Hampstead, NC (34E 23' N, 77E 41' W) - An 18 hole bermudagrass golf course with varied topography and soils. This was the northern-most sampling site in our study. Although the site had a fairly low population of mole crickets (primarily S. borellii) it was included in the study to extend the northern range of the study. This site was sampled weekly in 1996.

Oak Island Golf and Country Club (OI), Southport, NC (33E 54' N, 78E 04' W)- An 18 hole golf course with bermudagrass fairways and fairly uniform sandy soils located on Oak

Island in Long Beach. This site was sampled weekly in 1993 and 1996.

Oyster Bay Golf Links (OB), Sunset Beach, NC (33E 53' N, 78E 32' W)- An 18 hole golf course with bermudagrass fairways located on the mainland near Sunset Beach, just north of

74 the South Carolina border, and immediately adjacent to ST. This site represents the southern-most sampling site in this study. This site was sampled weekly in 1996.

Sea Trail Plantation (ST), Sunset Beach, NC (33E 54' N, 78E 31 W ) A complex of three

18 hole golf courses with bermudagrass fairways located on the mainland near Sunset

Beach, just a few miles north of the South Carolina border. Species composition and the level of infestation vary somewhat among the courses. This site was sampled weekly each summer from 1993–1997 and was also the site of a concurrent mole cricket flight study from

1995–1997. Two adjoining bermudagrass golf courses, Sandpiper Bay Golf and Country

Club (SP, two dates) and Angel Trace (AT, one date) were sampled in 1995 and 1997, respectively, to complete the year’s samples when sampling at ST was discontinued due to widespread insecticide treatment for mole crickets.

Nymph Sampling Procedure. Nymph sampling was conducted by using a standard sprinkling can to evenly distribute approximately 8 L of a 0.4% aqueous solution of dishwashing soap (Lemon Fresh Joy® ) to a 1.0 m2 area of turf (delineated by a 1.0 x 1.0 m

PVC frame). The solution acts as an irritant which causes mole cricket nymphs and other mobile soil arthropods to escape its effects by coming to the soil surface (Short and Koehler

1979). All emerging mole cricket nymphs were collected into 80% ethyl alcohol and preserved for later laboratory examination. Due to the extremely hot and dry conditions encountered in late July 1993, the sampling technique was modified. After 23 July 1993, all samples were made using twice the amount (16 L) of soapy water solution to flush the m2 area. Henceforth this will be referred to as the 2X soapy water flush (SWF) sampling method, and includes the majority of samples taken in the study. With a few exceptions, ten

75 samples were taken each week (n = 10) in 1993–1996. Twelve samples were taken each week (n = 12) in 1997. Three to four infested fairways were sampled at each site each week and usually no more than three samples were taken in a specific location. Sampling sites were randomly selected from within areas showing recent mole cricket damage, and therefore represent a judgement sample. On some occasions a separate set of five samples were randomly taken from adjacent undamaged areas.

Laboratory Processing of Specimens. The number, species, and mid-line pronotal length of all field-collected mole crickets were determined by examination under a dissecting scope equipped with an ocular micrometer. Variability in both size and the number of instars makes it impossible to accurately assign field collected nymphs to a specific instar (Hudson

1987, Braman 1993). Therefore, the pronotal length ranges presented by Matheny and

Stackhouse (1980) were used to categorize nymphs into either six (S. vicinus) or seven (S. borellii) size classes (Table 1).

Soil Moisture Sampling Procedure. Soil moisture samples were taken at all sites on each nymph sampling date during the 1994, 1995, and 1996 sampling seasons. Samples were taken from within the damaged areas in close proximity to the SWF sampling site on each fairway at the same time the m2 flushes were performed. Soil moisture was monitored first at two depths (1994), then at three depths (1995, 1996) by taking five random 31 cm (12- inch) deep soil cores using a standard soil sampling probe (2.4 cm dia). The top cm of grass and thatch was removed and each core was separated into portions representing 0–10.2,

10.2–20.3, and 20.3–30.5 cm in depth in the soil profile. The composite sample of cores from each depth range were bagged separately in plastic bags and chilled to prevent

76 moisture loss. The samples were returned to the laboratory and percentage soil moisture

(based on dry soil weight) was determined gravimetrically as described by Kramer (1969).

Measurement of Degree-days and Other Weather Data: Daily and monthly rainfall data from one National Weather Service station (Wilmington WSO Airport (UCAN 14407)) and three Extension Service weather stations (Southport 5N (UCAN 14354)), Shallotte

(SHAN7) and Longwood (UCAN 14192)) were obtained from the State Office of Climate at

North Carolina State University. Biophenometers (model TA51-P, Omnidata International,

Inc., Logan, Utah) that monitor temperature and automatically calculate degree-day accumulation at 10 minute intervals were installed at four locations. These sites were chosen to represent the northern, southern and central portions of the study area. The southernmost unit was on the driving range at ST. The northernmost site was LF and units were also located at FS and OI. Each unit was equipped with a soil temperature probe buried at a depth of 10 cm in an area representing average turf. The physiological limits of mole cricket development have not been determined, so each unit was programmed to accumulate degree-days within the range of 10–43EC, representing the generalized upper and lower thermal limits of insect development. Degree-day readings were recorded by the samplers at each sample date or course visit, and at other times by Master Gardener volunteers or golf course personnel. The units were put in place near the first of the year

(1993) and were checked daily to weekly throughout the sampling season.

Data Analysis. Total nymph counts, and counts for each size class were quantified as the mean number of nymphs per square meter for each species, site and date. Season-long class counts of less than 15 nymphs were excluded from further analysis. The date of the highest

77 mean soapy water flush count for each size class was identified as the peak date of abundance for that size class. Where a count of equal magnitude occurred on more than one date (ST94), the first date was considered as peak for this analysis. The quartiles (25, 50 and 75%) of peak date for the two smallest size classes (C1 and C2) were determined

(PROC UNIVARIATE) and examined by ANOVA (PROC GLM). Nymph count data for each size class were also used to make parametric estimates (PROC REGLIFE, logistic distribution) of the date at which 25, 50 and 75% cumulative abundance were reached. The three quartile dates for peak and cumulative abundance of C1 and C2 nymphs for each species were analyzed separately by ANOVA (PROC GLM) using a model with terms for

YEAR, SITE (YEAR), SPECIES, and YEAR*SPECIES, and another model with terms for

YEAR, SPECIES, and YEAR*SPECIES.

Julian date estimates for 25, 50 and 75% cumulative abundance were compared to corresponding cumulative soil degree-day estimates for those dates, and with an independent set of degree-day accumulations at twelve dates representing 10 day intervals

(10 April - 29 July) by correlation (PROC CORR). Where degree-day estimates were required at dates where readings were not available, fitted estimates made from and including the original data were used (PROC ILM). At sites where site-specific degree-day data were not available, estimates from the nearest biophenometer recording site were used.

Examination of differences in degree-days among years was made using a data set containing only site-specific degree-days.

Estimates of date for 25, 50 and 75% cumulative abundance were compared to monthly rainfall accumulations from the closest of the four weather stations mentioned above (PROC

78 CORR). Correlations between soil degree days, rainfall accumulations and the quartiles of abundance for C1 and C2 nymphs of each species were investigated further using models including various combinations of both independent climatic variables (PROC REG, PROC

GLM). All estimates of dates and degree-days, and all tests of equality, correlation, and analysis of variance were performed using SAS software (SAS Institute, 2000)

RESULTS

Sampling over all years and sites yielded a total a total of 14,753 nymphs composed of

9604 (65.1%) S. vicinus and 5149 (34.9%) S. borellii. When separated into size classes based on pronotal length, S. vicinus nymphs over all years and sites were represented by

18.2, 32.0, 19.9, 14.7, 9.9, 3.8 and 1.5% of the species total for size classes C1–C7, respectively. S. borellii nymphs over all years and sites were represented by 38.1, 30.6,

22.6, 7.0, 1.2 and 0.4% of the species total for size classes C1–C6, respectively.

Species counts from individual m2 samples were not recorded in 1993, however, nymph counts from single m2 samples during 1994–1997 ranged from 0–104, 0–68, and 0–105 for

S. vicinus, S. borellii, and both species combined, respectively. Mean sample data were available for all years and sites. Mean S. vicinus nymph density ranged from 0–29.7 (OB96) nymphs per m2. Mean S. borellii nymph density ranged from 0–24.4 (FS95) nymphs per m2.

Mean combined density of both species ranged from 0–30.5 nymphs per m2, with the maximum density occurring at the same year, date and site (OB96) as the highest mean density for S. vicinus nymphs. The date and magnitude of the highest mean combined density of both species at a site was strongly influenced by the date and magnitude of the highest m2 nymph count of the predominant species at the site.

79 Species Composition of Nymph Samples. Most of the sites were of mixed species composition, ranging from 0.8–97.3% S. borellii. Not all sites were sampled all years, precluding analysis of species composition among sites and years. However, annual data from two sites (FS and ST) are available, and indicate significant differences in species composition of the samples among years (Table 2). The magnitude of changes in species composition at these two sites precluded the analysis of both species at these sites in the later years of the study.

First Detection of Nymphs. Date of first detection for C1 nymphs ranged from 3 June

(FS96, ST96) to 1 July for S. vicinus and from 3 June (DH94) to 23 July (ST93) for S. borellii. C1 nymphs should have been the first and most numerous size class appearing in the samples, however, that was not always the case. C2 nymphs of S. vicinus were detected before the C1 nymphs on only one occasion (ST95) and on the same date as C1 nymphs on

13 (13/16= 81.3%) occasions. The C3 nymphs of S. vicinus were also detected before the

C1 nymphs on only one occasion (ST95), but occurred in the samples on the same date as the first C1 nymphs on 8 (8/16= 50.0%) occasions. Similarly, the C2 nymphs of S. borellii were detected before the C1 nymphs on only one occasion (ST94) and on the same date as

C1 nymphs on 9 (9/18 = 50.0%) occasions. C3 nymphs of S. borellii occurred in the samples on the same date as the first C1 nymphs on only one occasion.

The total numbers of C2 nymphs collected over the course of the season often exceeded the total numbers of C1 nymphs collected. The differences between C1 and C2 counts were generally less for S. borellii than for S. vicinus. The number of C2 nymphs exceeded the number of C1 nymphs in (15/16 =) 93.8% and (6/18=) 33.3% of the season-long samples

80 where S. vicinus and S. borellii were recovered, respectively. Over all years and season- long sampling sites, C1 and C2 nymphs represented 19.0 and 33.5% of the total nymphs collected for S. vicinus, respectively. C1 and C2 nymphs represented 38.6 and 30.9 % of the total nymphs collected S. borellii, respectively.

C1 and C2 size class counts used for analysis of peak count and cumulative abundance, and examination of degree-day and rainfall relationships ranged from 15–448 with a mean of 150 ± 113 (SD). Counts of 25 or more nymphs represent 88.9% of the data used. Sites were used as replicates for these analyses and comparison of sites was not meaningful due to the degree the set was unbalanced. Site within year effects were not significant when included in the models. Because the P-values of these effects substantially exceeded 0.20, site within year effects were pooled with error in the analyses.

Peak Count Date. Analysis of peak mean m2 counts included data for 15 site-years for each species. Peak count date represents the highest annual mean count for a species at a site, and is not necessarily statistically higher than mean counts on other dates. Examination of count data for S. borellii show that peak mean sample numbers of C1 and C2 nymphs occurred on the same date at a site in six of the 11 cases (54.5%) where the two classes could be compared. In the five cases where peak numbers of C1 and C2 were not found on the same date, peak C2 nymph numbers were either found the week after the peak C1 counts

(3 cases), at the next sampling date where sampling was not weekly (CF93), or occurred on the same date as the second peak number count for C1 nymphs at the site (FS94). Analysis for S. vicinus show that peak mean sample numbers of C1 and C2 nymphs occurred on the same date in 10 of the 13 cases (76.9%) where the two classes could be compared. Peak C2

81 nymph numbers were found one week after the peak C1 counts in the other three cases.

Peak C3 nymph counts occurred on the same date as peak C1 counts on two and five occasions for S. borellii and S. vicinus, respectively.

Estimates for the quartiles of peak mean count date for C1 and C2 nymphs are reported in Tables 3 and 3a. Analysis (ANOVA) of peak count date for C1 nymphs included data for

13 and 14 site-years for S. vicinus and S. borellii, respectively. Results indicate that year effects were not significant for either S. vicinus (df = 4; F = 2.88; P = 0.0950) or S. borellii

(df = 4; F = 0.04; P = 0.9958), however, there was a highly significant difference in peak date between species for C1 nymphs (df = 1; F = 16.28; P = 0.0009). Peak mean count date for C1 nymphs ranged from 21 June - 2 July (Julian date (JD) 172–183) and 9–13 July (JD

190–194) for S. vicinus and S. borellii, respectively. Mean peak count dates for C1 nymphs of S. vicinus and S. borellii were 25 June (JD 175.9 ± 2.7) and 11 July (JD 191.8 ± 2.9), respectively. The timing of peak mean counts for C1 nymphs are presented in Figure 1.

Analysis of peak date for C2 nymphs (n = 15 and 12 site-years for S. vicinus and S. borellii, respectively) gave similar results indicating that year effects were not significant for either S. vicinus (df = 4; F = 1.70; P = 0.2254) or S. borellii (df = 4; F = 1.21; P = 0.3861).

As in the C1 analysis, the species effect for C2 peak date indicate a highly significant difference in C2 peak date between species (df = 1; F = 36.46; P < 0.0001). Peak count dates for C2 nymphs ranged from 20 June - 2 July (JD 171–183) and 9–25 July (JD

190–206) for S. vicinus and S. borellii, respectively. Mean peak date for C2 nymphs of S. vicinus and S. borellii was 26 June ( JD 177.0 ± 2.2) and 16 July (JD 197.1 ± 2.5), respectively. The timing of peak mean counts for C2 nymphs are presented in Figure 2.

82 Examination of peak count data indicate that mean C1 peak count date of S. vicinus occurred

16 d earlier than the peak for S. borellii. Peak count data for C2 nymphs indicate that mean peak date of S. vicinus occurred 20 d earlier than peak for S. borellii. The ANOVA results for the species*year interaction for both C1 (df = 4; F = 0.33; P =0.8510) and C2 nymphs

(df = 4; F = 2.10; P =0.1254) indicate the relative difference in peak date between species was approximately the same among years. The difference in mean dates for C1 nymphs indicate an 11, 15, and 23 d difference between species for 25, 50, and 75% abundance, respectively. The difference in mean dates for C2 nymphs indicate a 20, 24, and 24 d difference between species for 25, 50, and 75% abundance, respectively. Averaging the differences in date among quartiles returns values similar to those suggested by the differences in mean timing between species for both size classes (16 and 23 d difference between species for C1 and C2 nymphs, respectively).

Observations on Peak Counts. Examination of mean C1 and C2 nymph counts indicate that there were two distinct peaks for hatch for both species in some cases. A count was considered as a second peak if it occurred at least two weeks after the first peak date, was numerically greater than the intervening counts, and equaled at least 30% of the initial peak count. With one exception (ST94), the second peak was of smaller magnitude, often representing less than half the number C1 or C2 nymphs per m2 observed in the first or primary peak. However, peaks of similar magnitude were observed at one site (ST). Data for S. vicinus indicate that a second peak occurred in four out of 13 cases (30.8%) for C1 and six out of 15 cases (40.0%) for C2 nymphs. Data for S. borellii indicate that a second peak occurred in only two out of 12 cases (16.7%) for C1 and one out of 12 cases (8.3%) for

83 C2 nymphs. The occurrence of more than one peak in mean nymph counts suggest that analysis of cumulative count data was necessary in order to more fully characterize hatch.

Cumulative Nymph Abundance. Analysis of date for cumulative abundance of C1 nymphs include data for 13 and 14 site-years for S. vicinus and S. borellii, respectively.

Similar analysis for C2 nymphs include data for 15 and 12 site-years for S. vicinus and S. borellii, respectively. Parametric estimates of date for 25, 50 and 75% abundance of C1 and

C2 nymphs were examined for both species together and separately. Estimates of date for the quartiles of C1 and C2 nymphs are reported in Tables 4 and 4a. Sampling started later in

1993 and continued later in 1993 and 1994 than in other years. The number of sites, extended season, non-weekly sampling period at some sites (CF), and increased sampling period in late season (at weekly sampling sites), contributed to greater variability observed in all three date estimates for S. borellii C1 nymphs in 1993 as compared to other years. Use of the 1X SWF, and switch to the 2X sampling method in late season may also have contributed to increased variability. Variability in C2 date estimates was most pronounced for S. borellii at 25% and S. vicinus at 75% cumulative abundance than for other quartiles.

ANOVA results for date of abundance for S. vicinus C1 nymphs show significant differences among years for each quartile of abundance (df = 4; F = 4.31; P =0.0376 for

25%; F = 9.34; P = 0.0042 for 50%; F = 8.04; P =0.0066 for 75%). Conversely, analysis of date for S. borellii C1 nymphs show no significant differences among years for any quartile of abundance (df = 4; F = 0.25; P = 0.9003 for 25%; F = 0.32; P = 0.8576 for 50%; F =

0.40; P = 0.8073 for 75%). When both species were included in the model, there were highly significant differences in timing between species for all three C1 date quartiles (df =

84 1; F = 24.48; P =0.0001 for 25%; F = 27.44; P < 0.0001 for 50%; F = 26.00; P < 0.0001 for

75%).

Results of the analysis of date for S. vicinus C2 nymphs show significant differences among years at 25% (df = 4; F = 5.13; P = 0.0165), near significance at 50% (df = 4; F =

3.45; P = 0.0512), but no significant difference in date among years for 75% (df = 4; F =

2.26; P = 0.1346) cumulative abundance. Similar analysis of date for S. borellii C2 nymphs show no significant differences among years for either 25% (df = 4; F = 0.82; P = 0.5516) or 50% (df = 4; F = 2.48; P = 0.1393) cumulative abundance, however, results at 75% (df =

4; F = 5.85; P = 0.0216) indicate significant differences in date among years. As seen with the two species analysis of C1 nymphs, analysis of C2 quartile dates also indicate there were highly significant differences in the timing between species (df = 1; F = 47.26; P < 0.0001 for 25%; F = 50.27; P < 0.0001 for 50%; F = 36.79; P < 0.0001 for 75%). Analysis of the species*year interaction for both C1 (df = 4; F = 0.43; P =0.7849 for 25%; F = 0.40; P =

0.8088 for 50%; F = 0.35; P = 0.8432 for 75%) and C2 nymphs (df = 4; F = 0.65; P =0.6369 for 25%; F = 0.76; P = 0.5674 for 50%; F = 0.78; P = 0.5527 for 75%) indicate the difference in timing between species was approximately the same from year to year. Mean date for each quartile of abundance for C1 and C2 nymphs of both species are reported in

Table 4. The difference in mean quartile dates indicate a 15 d difference between C1 nymphs, and a 16 d difference between C2 nymphs of the two species.

Correlation of quartile date and corresponding soil degree-days. Analysis of soil degree-days and cumulative abundance of C1 nymphs included data for 11 site-years for both S. vicinus and S. borellii. Similar analysis for C2 nymphs included data for 13 and 10

85 site-years for S. vicinus and S. borellii, respectively. All three quartile dates for both size classes and species were correlated with their respective soil degree-day accumulations at that date. Mean soil degree-days (± SD) for each quartile are reported in Tables 5 and 6.

Analysis for S. vicinus (n = 11 site-years) show that the dates of 25, 50 and 75% cumulative abundance of C1 nymphs are correlated with soil degree-day accumulations to that date

(Pearson Correlation Coefficient (PCC) = 0.62, P = 0.0424 for 25%; PCC = 0.69, P =

0.0195 for 50%; PCC = 0.82, P = 0.0019 for 75%). Quartile dates for S. vicinus C2 nymphs

(n = 11 site-years) are highly correlated with their corresponding soil degree-day accumulations (PCC = 0.89, P < 0.0001 for 25%; PCC = 0.92, P < 0.0001 for 50%; PCC =

0.95, P < 0.0001 for 75%). Soil degree-days for S. vicinus C1 nymphs ranged from

1009–1196, 1116–1374 and 1241–1557 for 25, 50 and 75% cumulative abundance, respectively. Soil degree-days for S. vicinus C2 nymphs ranged from 945–1361, 1072–1575 and 1214–1768 for 25, 50 and 75% cumulative abundance, respectively.

Analysis for S. borellii C1 nymphs (n = 11 site-years) show that the dates of 25, 50 and

75% abundance were highly correlated with their corresponding soil degree-day accumulations (PCC = 0.93, P < 0.0001 for 25%; PCC = 0.94, P < 0.0001 for 50%; PCC =

0.92, P < 0.0001 for 75%). Quartile dates of abundance for S. borellii C2 nymphs (n = 10 site-years) were also highly correlated with their corresponding soil degree-day accumulations (PCC = 0.91, P = 0.0002 for 25%; PCC = 0.89, P = 0.0006 for 50%; PCC =

0.89, P = 0.0005 for 75%). Soil degree-days for S. borellii C1 nymphs ranged from

1222–1730, 1339–1827 and 1432–1904 for 25, 50 and 75% cumulative abundance, respectively. Soil degree-day accumulations for S. borellii C2 nymphs ranged from

86 1346–1808, 1408–1887 and 1472–1962 for 25, 50 and 75% cumulative abundance, respectively.

Degree-day accumulations and cumulative abundance may be correlated because of the obvious correlation between degree-days and date. Later dates are likely to have greater degree-day accumulations than earlier dates. Therefore, differences among years were further examined using degree-day accumulations at uniform dates.

Relationship Between Nymph Abundance and Soil Degree-Days. Dates for the three quartiles of C1 nymph abundance were compared with soil degree-day accumulations at twelve dates representing 10 d intervals (10 April - 29 July; JD 100–220) prior to and during the nymph sampling period each year to determine if differences in degree-days explain variation in timing (date) among years. The dates of 25 and 50% cumulative abundance for

C1 S. vicinus were significantly correlated with soil degree-days at six and four dates, respectively (Figure 3) with maximum correlation at 30 May. The date of median C1 S. vicinus nymph abundance is 27 d after 30 May. No significant correlations between soil degree-days and date of 75% abundance of C1 S. vicinus were found at any date, however, a similar trend of the maximum value of PCC (-0.58719; P = 0.0575) occurring on 30 May was observed. No significant correlations between soil degree-days and quartile dates for S. vicinus C2 nymphs were found at any date tested. However, a similar trend in PCC values among test dates was noted with maxima at 19 June (JD 170), or 20 d later than the maxima for C1 nymphs. Similar analysis for S. borellii C1 and C2 nymphs show no significant correlations between soil degree-days and any quartile of abundance.

Soil degree-day accumulations used for correlation analysis of C1 S. vicinus at 30 May

87 ranged from 676–855 with a mean of 778.9 ± 66.1 (SD). Soil degree-day accumulations used for correlation analysis for C1 S. borellii nymphs at 30 May ranged from 672–855 with a mean of 751.1 ± 63.7 (SD).

Soil degree-day accumulations at 30 May (just prior to hatch) show the most significant correlations with nymph abundance and were examined further using ANOVA. Results indicate that site-specific accumulations were significantly different among years for S. vicinus C1 nymphs (df = 4; F = 39.59; P = 0.0018), but not for S. borellii C1 nymphs (df =

3; F = 8.60; P = 0.0553). Although inclusion of degree-days in various models partially explain variation among years for S. vicinus in all cases examined, the most compelling and significant relationship is the analysis for date of 50% C1 nymph abundance. The addition of 30 May closest site degree-day accumulations to this model explain significant differences in date among years (df = 4; F = 6.16; P = 0.0360) by reducing the observed variation (df = 4; F = 2.30; P = 0.1924).

Relationship Between Nymph Abundance and Rainfall. Correlation analysis of rainfall and cumulative abundance of C1 nymphs included data for 13 and 14 site-years for S. vicinus and S. borellii, respectively. Similar analysis for C2 nymphs included data for 15 and 12 site-years for S. vicinus and S. borellii, respectively. Date estimates for 25, 50 and

75% cumulative abundance for C1 and C2 nymphs were compared to monthly rainfall accumulations for the preceding April, May and June (when the majority of eggs are laid).

A significant correlation was found between rainfall in June and date for all three quartiles of abundance for C1 S. vicinus nymphs (PCC = -0.69916; P = 0.0078 for 25%; PCC = -

0.72277; P = 0.0053 for 50%; PCC = -0.64062; P = 0.0183 for 75%). Additionally, rainfall

88 during a two week interval from 27 May to 9 June was similarly correlated with all three quartiles of abundance for C1 S. vicinus nymphs (PCC = -0.57782; P = 0.0386 for 25%;

PCC = -0.67043; P = 0.0121 for 50%; PCC = -0.67012; P = 0.0122 for 75%). No significant correlations between rainfall and cumulative abundance of S. borellii C1 nymphs were found at any date.

Date of cumulative abundance for both S. vicinus and S. borellii C2 nymphs were significantly correlated with rainfall in both May and June. All three quartiles dates for S. vicinus C2 nymph abundance were correlated with rainfall in May (PCC = -0.64131; P =

0.0100 for 25%; PCC = -0.70499; P = 0.0033 for 50%; PCC = -0.74295; P = 0.0015 for

75%). Only the date of 25% cumulative abundance was correlated with rainfall in June

(PCC = -0.55405; P = 0.0321). Date of 50% cumulative abundance for C2 S. borellii nymphs were significantly correlated with rainfall accumulations in both May (PCC = -

0.62111; P = 0.0311) and June (PCC = -0.66316; P = 0.0187). Date of 75% cumulative abundance for C2 S. borellii nymphs was also significantly correlated with rainfall in both

May (PCC = -0.70747; P = 0.0101) and June (PCC = -0.78171; P = 0.0027). Correlations with rainfall in May and June may be due to an effect on oviposition or hatch, however, correlation with June rainfall could be due to differential SWF efficacy effects resulting from differences in soil moisture. Correlation with rainfall at the end of May and early June appear to be related to oviposition.

Models Including Both Soil Degree-Days and Rainfall. Models (GLM) including soil degree-day accumulations at 30 May (JD150) and monthly rainfall accumulations (April,

May and June) were examined to determine if inclusion of both climatic variables explained

89 variation in date of 25, 50 and 75% C1 nymph abundance better than either variable alone

(Table 7). Although both soil degree-days and rainfall were correlated with S. vicinus C1 nymph abundance, the inclusion of both independent variables in models only explain

2.1–16.9% more variation in date than either variable does independently. The r2 values indicate that 30 May soil degree-days and June rainfall together only explain 65.8, 64.4 and

48.6% of the variation in date of 25, 50 and 75% S. vicinus C1 abundance, respectively.

Additionally, these variables are not significant in the models when both are included. Both variables explain similar amounts of variation in median date of abundance (Table 8), however, degree-days were only a significant source of variation at 25 and 50% abundance.

June rainfall was significant and had negative parameter coefficients for all quartile dates.

Similarly, the r2 values indicate that 30 May soil degree-days and May rainfall together explain 52.9, 58.6 and 60.3% of the variation in date of 25, 50 and 75% S. vicinus C2 abundance, respectively. This is only 5.1–11.8% more variation than is explained by rainfall or degree-days alone. No significant correlations were found with either degree- days or rainfall for any quartile of S. borellii C1 nymph abundance, therefore it is not surprising that we failed to find a significant model including both variables.

DISCUSSION

The goal of the research was to gather information useful for the prediction of oviposition and hatch for the two pest mole cricket species. This required the use of a repeatable, non-destructive technique to sample the nymph population, a method to classify both species of nymphs according to size or instar, and the collection of climatic data.

Observations on Sampling Techniques. As soil-dwelling insects, mole crickets are

90 notoriously difficult to observe or sample with the main problem being that of relating the sample statistics to the true population parameters (Walker 1984, Hudson 1989). A variety of methods have been used to sample both nymphs and adults, including a tractor mounted soil corer (Williams and Shaw 1982), tree spade sampling (Walker 1984), linear pitfall traps

(Lawrence 1982), surface counts of pesticide-treated areas (Short and Koehler 1977, Walker

1979), irritant soil drenches (Walker 1979, Short and Koehler 1979), sound traps for adults

(Walker 1982), and a rating system to relate surface damage to population (Cobb and Mack

1989). However, all methods yet proposed have their limitations. Sampling in the current study was conducted on golf courses where it was not possible to use turf-damaging techniques, which limited the selection of methods. The use of insecticide drenches for sampling was impractical due to concern for the safety of the samplers, golfers, golf course personnel, and the environment. Surface counts of pesticide-treated areas were also considered an unreliable measure of the population due to the possibility of differences in species and instar susceptibility. Additionally, a large proportion of the population (50%) dies below the surface in insecticide or bait-treated areas (Walker 1979). Although Cobb and Mack (1989) quantified the relationship between damage ratings and population density, this technique can only be used when damage is visible, and can not be used to directly quantify development of the smaller instars.

Soapy Water Flush Sampling Method. The objectives of the research required the collection of an adequate number of nymphs to characterize the timing of hatch, and greatly influenced the selection of the sampling program used here. The soapy water flush (SWF) technique of Short and Koehler (1979) was selected because it is easy, safe, practical,

91 inexpensive, does not harm the turf, delivers immediate results, and yields large numbers of good quality specimens for species determination and size classification.

Many factors affect the efficacy of the SWF sampling method. The depth to which the flush solution penetrates the soil varies with soil moisture, soil type, compaction, amount of turf coverage, and a host of other factors. The effects of soil moisture on flush efficacy have been documented by Hudson (1989). Working with a confined population, he found that flushing efficiency was a function of soil moisture varying from only 50% efficiency at

13.3% to 90% efficiency at 19.5% soil moisture. Soil moisture at our sites was a function of both irrigation and rainfall, and varied substantially among sampling dates and sites. The depth and population density of the nymphs can also vary tremendously. Nymphs avoid hot or dry conditions near the surface, and can retreat even deeper as they grow and expand their tunnels into the soil profile. Spatial variation in distribution also occurs as the nymphs grow and spread out laterally from the area of hatch into adjacent turf. Although the timing of daily activity has never been quantified, temporal variations due to feeding and tunneling must also occur and affect sampling results.

Size is an important factor for two reasons. Early in the season the nymphs are small and often come to the surface before the soap solution has been entirely applied. This indicates that they are near the soil surface, and unable to escape the irritant flush. The flush technique is probably most effective at this time. As summer progresses, the nymphs grow and are able to move faster and tunnel to greater depths, enabling them to better avoid the drench. This is most evident in the fall when it is difficult to get large nymphs and adults to the surface. Additionally, large nymphs seem less affected by the irritating effects of the

92 solution. This is obvious when attempts are made to keep nymphs alive after they come to the surface. Even when they are washed off immediately, small nymphs rarely survive while a much larger proportion of large nymphs recover from the effects. This suggests that some small nymphs may succumb to the effects of the drench before they reach the surface.

The drop in sample numbers in late July 1993 was probably the result of both size and dry soil conditions, and precipitated the increase in sample flush volume. The decision to double the quantity of soapy water was unfortunate with regard to consistency that year, but served to meet the overall objectives of the sampling in later years. Extremely dry soil conditions reduce penetration of the flush and cause the crickets to move deeper in the soil.

Doubling the volume of drench may not result in substantially higher numbers of small nymphs under normal moisture conditions, but almost certainly increases the yield of nymphs in dry weather. It also increases the yield of larger instars later in the season.

Examination of the sample data after the increase in flush volume shows that total nymph counts at FS more than doubled compared to the previous week, with the overwhelming majority of both species represented by C3 and C4 nymphs. We believe this is a direct result of adopting the 2X method combined with the prevalence of these size classes at this time of year.

It is reasonable to assume twice the volume of liquid irritant should provide greater flushing and penetration of the tunnel system and greater soaking and contact time in the soil column near the surface. However, the increase in volume also effectively increases the size of the sample unit as depth of penetration is the only dimensional variable of the sample. Of course, this also varies for each sample due to soil compaction, soil type, soil

93 saturation, thickness of the thatch layer, and the amount of tunneling by the crickets.

The change in the sampling procedure took place late enough to have a minimal affect on

S. vicinus C1and C2 nymph data because the majority of the nymphs had already transformed to larger instars. Count data at the sites used for analysis indicate that only 1.8

(5/284) and 5.4% (27/506) of the total number of C1 and C2 nymphs were taken in samples after changing to the 2X flush. Date estimates for S. borellii may have been more seriously affected. Count data at the sites used for analysis indicate that 15.5 and 43.5% of the total number of S. borellii C1 and C2 nymphs were collected after changing to the 2X flush.

Although this may partially be the result of the increased flush volume, at least some of these nymphs would have been collected using the original flush volume. These results are also a function of the later and extended hatch for this species combined with the extended sampling season that year. Weekly sampling continued well into August, with less frequent sampling continuing at some sites until the end of October. The increase in flush volume and the late sampling may have contributed to the greater variability and slightly later estimates for hatch than in years when sampling ended at the end of July. However, it should be noted that a few S. borellii continue to hatch well after the end of July. C1 nymphs of this species were found in our samples as late as 26 August in 1993, and have even been found as late as October. Therefore, these later estimates may quantify S. borellii hatch better than where sampling was discontinued at the end of July. These nymphs represent only a small proportion of the population and it has been suggested that late- hatching nymphs perish (Walker and Nation 1982) and should not be considered in the timing of management practices.

94 Sample Selection and Sampling Unit. Randomness is a desirable attribute of any sampling program, but is not always practical. Although the samples were taken randomly within areas of visible mole cricket damage, undamaged fairways were not sampled. Random sampling would have wasted much sampling time and led to the recovery of inadequate numbers of specimens. Therefore, the procedure must be defined as a judgment sample.

Jessen (1978) points out that the strongest case for judgment selection in sampling is where sample size is small, the sample universe (the fairway surface here) is visible to the sampler, and that the selector has proven skill in the selection. These criteria have been met here.

The m2 sample is probably the largest unit that can be easily replicated in the field by an individual sampler. It is small enough to examine fairly intensively yet large enough to ensure mostly non-zero counts from damaged areas. The ten samples took approximately three hours per site (20 min per sample on the average) including moving between sampling locations, hauling water and delays due to play. Although sampling multiple frames at once can shorten sampling time, it would be difficult to substantially increasing the number of samples while sampling multiple sites.

Species Composition and Population Density. Data on changes in species composition among years were included here to illustrate the magnitude and speed at which such changes can occur (Table 2). Although some of the differences in species composition among years may be due to natural cycles and phenomena, selection of insecticides, treatment date and other management practices were likely a significant factor influencing these changes.

Data on the highest counts per single m2 and highest mean number of nymphs per m2 were complied to demonstrate the range of nymph counts and densities encountered in

95 sampling, and are not necessarily statistically greater than other counts on the same or different dates. The highest counts in a single m2 did not always occur near peak hatch, and probably resulted from sampling in areas where multiple clutches had hatched. Although the magnitude of most of the highest mean nymphs per m2 are similar to reported clutch size, the distribution of counts among size classes indicate that the nymphs do not represent hatching of a single clutch.

Our observations and those of other authors (Kleyla and Dodson 1978, Walker 1979,

Hudson 1985, Fowler 1989) suggest that mole cricket populations have a highly clumped distribution. Increasing numbers in our samples from undamaged areas adjacent to the damage samples confirm that nymphs become more randomly distributed over the course of the season (unpublished data). The means and ranges of nymph densities reported here show the range of densities within infested areas and can not be considered as average course or fairway infestation. Additionally, results of SWF and all other non-destructive methods can not be considered as an absolute sample and must underestimate the true population density within the damaged areas.

Size Classification of the Nymphs. Mole crickets develop from egg to adult by gradual metamorphosis, increasing in size by molting through a variable number of instars (Hudson

1987, Braman 1993). The variability in both number and size of the instars make it impossible to definitively assign field collected nymphs to a specific instar. Nevertheless, the present study required the separation of nymphs by instar or size class in order to quantify hatch and development. Unfortunately, few studies have attempted to catagorize mole crickets nymphs by size, or relate size with instar. Walker (1979) used total body

96 length to assign nymphs of both species to six size ranges, from small to very large.

Although this may be adequate for some purposes, it was not useful in our attempt to delineate hatch and development. Variation in the length of the unsclerotized abdomen due to both feeding and preservation in alcohol can result in extreme variability in total length.

Only one study has related size to actual instar. Hudson (1987) followed the number of instars and pronotal length of S. borellii [as S. acletus] in a laboratory study using a small sample of nymphs (n = 24) reared from a single clutch of eggs. Only five of the original hatchlings developed to adults, with over 50% of the mortality occurring by the fourth instar. When the pronotal lengths and associated standard deviations are used to produce size ranges (rounded to the nearest 0.1mm) for each instar, many of the ranges are not contiguous or overlap, making it impossible to assign our field-collected nymphs to one of the instars. Hudson (1987) noted that the size variability exhibited in nymphs from a single clutch of eggs suggested even greater variability in the field population. Additionally, the effect of the laboratory diet, temperature and moisture may have had a significant effect on growth.

Matheny and Stackhouse (1980) took a different approach to size classification. They used statistical techniques based on Dyer’s Law of geometric progression to fit a large sample of field-collected nymphs to seven size classes for S. vicinus (n = 2068) and six size classes for S. borellii (n = 605). Because their study included large sample sizes of field- collected nymphs for both species, and used uniform statistical techniques to delineate size classes, we used the pronotal size ranges of Matheny and Stackhouse (1980) to classify nymphs in the present study. The fact that the actual number of instars is greater than the

97 number of size classes presented by Matheny and Stackhouse is unimportant here, as classification of only the smaller instars were necessary to quantify hatch. It should be noted that the pronotal range presented by Hudson (1987) for the first instar of S. borellii is in close agreement with the first size class (our C1) of Matheny and Stackhouse (1980), as is the lower range measurement for the second instar. This suggests that our field-collected first instars of this species were less likely to be classified as C2. However, Matheny and

Stackhouse’s upper range measurement for the second size class (our C2) much exceeds the lower range for Hudson’s third instar, suggesting that some third instars might have been classified as C2. Although this may help to explain the greater number of C2 than C1 nymphs for S. borellii, this species has two morphologically distinct forms (four-spotted and mottled), both of which have been introduced in multiple locations, and exist in disjunct populations (Walker and Nickle 1981). Giver that the nymphs in each study may have come from different populations, and that we are comparing laboratory results with those in the field, differences in size range might well be expected. Although there are no comparative size data for S. vicinus, size variation among different populations is also possible for this species, and could have resulted in first or third instar nymphs being classified as C2.

Therefore, it is impossible to further resolve size classification differences in either species.

Relative Numbers of C1 and C2 Nymphs. Although the newly-hatched first instar should clearly be the most abundant nymphs in the population, in most cases the C2 nymphs were the most numerous size class found in the samples. The number of C2 nymphs exceeded the number of C1 nymphs in (15/16) 93.8% and (6/18) 33.3% of the season-long samples for S. vicinus and S. borellii, respectively, indicating this phenomenon occurred almost three times

98 more often with S. vicinus than S. borellii. There are several plausible explanations for these results. As discussed above, it is possible that some first or third instar nymphs were inadvertently classified as C2 nymphs due to the size ranges used, thereby increasing the relative number of C2 to C1 nymphs. It is also possible that the sampling method was more effective on C2 than C1 nymphs because not all C1 nymphs were close enough to the surface or able to emerge in response to the irritant solution. As the smallest size class, C1 nymphs are likely less able to dig to the surface after tunnels are collapsed by the flush solution and may succumb to the flush while still in the soil. Due to their small size, those that do make it to the surface are more difficult to find than the larger size classes.

Additionally, the relatively small numbers of samples taken per sampling date are probably not sufficient to detect the first clutches that hatch, further reducing the total number of C1 nymphs collected. This hypothesis is supported by the fact that C2 nymphs were first detected on the same date as C1 nymphs 81.3 and 50.0% of the time for S. vicinus and S. borellii, respectively. It has also been reported that some C1 nymphs do not immediately exit the egg-chamber. This may also reduce the yield of C1 nymphs, a possibility that is discussed below in the section on hatch date. As discussed in the section on sampling above, low soil moisture condition can also reduce the number of nymphs collected. Where dry conditions occur in early season the yield of C1 nymphs may be reduced. This would mainly affect collection of the earlier-hatching S. vicinus nymphs. Any combination of these factors is also possible.

The similar date of first detection and the relative proportion of C1 and C2 nymphs indicate this phenomenon occurred more frequently with S. vicinus than S. borellii,

99 suggesting a differential effect between species. However, not all circumstances would have caused differential results. For example, difficulty of the samplers in finding the small C1 nymphs should have affected both species equally. The same reasoning holds true with regard to the possibility that C1 nymphs succumb to the drench before surfacing, although a differential effect between species cannot be ruled out. However, both low soil moisture in

June and the previously mentioned difficulties with size classification could have disproportionally affected S. vicinus C1 counts. If low soil moisture effects on sampling efficiency were the cause, we would expect the effect to be more pronounced in some years

(or sites) than others. This was not the case. S. vicinus C2 numbers exceeded C1 numbers fairly uniformly across sites and years. Without further data there is no reliable way to choose among the alternative explanations above, however, both size classification and attributes of the sampling method favoring the recovery of larger instars over the first instar seem to be the most viable hypotheses.

Because of the disparity between total numbers of C1 and C2 nymphs, and low counts that made it impossible to quantify hatch using C1 numbers in some cases, we attempted to quantify development using both size classes. Similar use of larger size classes was not considered because of the increasing error factor introduced by extending the range of development time, the decreasing effectiveness of the sampling technique, and ending sampling before the completion of these instars in the field. As the results of analysis produced little differences between estimated dates for C1 and C2 nymphs, and the possibility that some first or third instars were classified as C2, use of the C1 data is more appropriate to quantify the timing of these events. Although we would have preferred to

100 limit the use of nymph counts as low as 15 in analysis of date, such counts represent a small percentage of the total data used. Examination of date estimates suggest these estimates were in line with higher count date estimates and limited the loss of site-years.

Peaks in C1 and C2 Nymph Counts. Although two or more peaks in counts were clearly evident in some cases, defining the phenomenon proved difficult. Therefore, both the number of cases where this occurred, and the time interval between peaks are open to interpretation. The phenomenon may simply be a sampling artifact due to sampling in areas of variable population density on different dates, or a manifestation of differences in sampling efficiency due to variability in soil moisture levels. Given the small numbers of samples per sample date (10 or 12), both explanations are plausible. Rainfall data definitely indicate increased soil moisture levels prior to some second peak events, however, the higher counts may have resulted from either an increase in sampling efficiency (Hudson

1989) or an effect on hatch or escape from the egg chamber. However, alternative explanations are also offered by the literature.

Both species are known to deposit eggs in up to ten separate clutches (Forrest 1986,

Hudson 1987) with number of eggs per clutch generally decreasing with each successive clutch (Braman 1993). Additionally, Ngo and Beck (1982) attributed cyclic12-d peaks in flight activity to female oviposition cycles and Forrest (1986) found 7–12 day oviposition cycles between successive clutches for S. vicinus and 8–11 day cycles for S. borellii.

Regardless of the duration of time between clutches, most females do not lay their entire complement of eggs at one time, and multiple clutches must have multiple hatch dates. The existence of more than one peak in C1 and C2 nymph counts in the field may be evidence of

101 this phenomenon. Progressively decreasing clutch size (Braman 1993) and reduction in percentage hatch as the season progresses (Forrest 1986) might also explain the lesser magnitude of the second peak in nymph counts.

Another factor that could explain the occurrence of more than one peak in hatch is the existence of two distinct types of development described by Hayslip (1943). Hayslip observed that egg deposition in Florida began about one month earlier for adults that transformed to the adult stage in the fall than those that molted to the adult stage the following spring. On some occasions where two peaks occurred in C1 nymph counts in NC, the interval between peaks was approximately one month, which agrees with the Hayslip hypothesis, however, this interval was not observed in all cases and the magnitude of the peaks does not seem to agree with the proportion of the population that transform in the spring and fall. However, Forrest (1986) found a significant positive relationship between the size of female S. borellii and the mean number of eggs per clutch. This may lend support to the “two type” scenario, and explain the difference between the magnitude of the two peaks observed. Although not followed quantitatively, we have noted that S. borellii captured in early spring flights seem to be larger than those captured later in the summer.

This may be evidence of the differences in size between fall and spring adult transformation

(Forrest 1987) and may explain the difference in the magnitude of the peaks.

Date Estimates for Peak and Cumulative Abundance. Peak date estimates were made from a data set composed of one count date at each site each year. These counts may not be statistically different from counts made on other dates, and may have been affected by chance selection of high density sampling sites, variations in soil moisture affecting

102 sampling efficacy or other factors. The existence of more than one peak in nymph counts, and lack of distinct peaks in other cases suggest that it is more prudent to rely on our estimates of cumulative abundance. Cumulative abundance data take into consideration all the nymphs taken over the entire season, and should provide more reliable estimates.

Using Nymph Counts to Determine Hatch Date. In order to make estimates for oviposition or hatch from our data, we must equate the dates of C1 nymph abundance with hatch date. However, it is likely that the SWF technique used here would only flush up nymphs that had previously escaped the sealed egg chamber, and were near the surface.

Unfortunately, it is not clear when newly-hatched mole cricket nymphs leave the egg- chamber and arrive at the surface. Van Zualuwenburg (1918) reported that newly-hatched

Scapteriscus nymphs remain together in the egg chamber for a few days, and Hayslip (1943) reported it took 9 days for S. vicinus nymphs to escape the egg-chamber and tunnel to the surface under laboratory conditions. Worsham and Reed (1912) also report that S. vicinus

[as S. didactylus] nymphs remain in the egg-chamber for the first few days, presumably an observation from cage studies. Forrest (1986) reported that although the eggs in a clutch hatch within a 24-hour period, it is a few days until the young tunnel to the surface. If such a delay exists, it would have important implications for the current study because it would effectively halve the time that detection of the first instar was possible by the flush technique. This would suggest that the C1 nymphs we equate with the first instar may have already proceeded through half their duration of existence prior to the collection date, skewing all estimates using those dates. Because the previously discussed differences in the relative numbers of C1 and C2 nymphs (especially for S. vicinus) could have resulted from

103 this behavior, we feel it important to mention this possibility, and the uncertainty it introduces into the hatch estimates. However, for the purposes of timing sampling or treatment, the actual samples taken represent the population that would have been subjected to chemical or biological agents, had treatment taken place on that date. Estimates for date of oviposition are helpful in delineating hatch, however, abundance of small nymphs is the most important factor when this stage is the target for control.

Estimates of duration for hatch and development: Egg development time from oviposition to hatch is a function of temperature, and may be affected by moisture as well.

Estimates of the time required for the eggs to hatch and the duration of each instar for the two species included here are reported by only a few authors. Hayslip (1943) presented data on egg deposition and incubation time (March through September) for both species from a field cage study in Florida. Mean incubation period over the entire egg-laying period was

18.5 and 21.0 d for S. vicinus and S. borellii, respectively. However, the incubation period varied substantially according to the month the eggs were deposited, ranging from 11.0–32.0 d for S. vicinus and 16.1–37.0 d for S. borellii. The highest percentage of eggs for both species were laid in May with an incubation time of 17.9 and 22.0 d for S. vicinus and S. borellii, respectively. Over half of all eggs for both species were laid by 1 June with 98.9 and 89.5% of the eggs laid by 1 July for S. vicinus and S. borellii, respectively. Worsham and Reed (1912) give an incubation range of 24–26 d for the eggs of S. vicinus [as S. didactylus] with egg deposition occurring between 15 April and 15 June in Georgia. These estimates fall within the overall range presented by Hayslip (1943) and suggest oviposition starts later in Georgia than in Florida.

104 Laboratory data presented by Braman (1993) show the range of egg development time at

27EC. was 11.4–20.0 d for S. vicinus and 17.3–24.7 d for S. borellii. Mean duration of the egg stage was 16.3 ± 0.5 (SE) and 20.9 ± 0.3 (SE) d for S. vicinus (n = 15) and S. borellii (n

= 52), respectively. Using a base temperature of 10EC. (50EF) this translates to 277.4 ± 7.9 and 356.1 ± 5.2EC. (499.4 ± 14.3 and 641.0 ± 9.4EF.) degree-days for the eggs of S. vicinus and S. borellii, respectively. Duration of the first instar ranged from 10.6–17.1 d for S. vicinus and 15.6–28.2 d for S. borellii. Mean duration of the first instar was 14.1 ± 0.7 and

21.9 ± 0.8 d for S. vicinus and S. borellii, respectively, translating to 239.4 ± 11.2 and 371.7

± 13.4EC (430.8 ± 20.1 and 669.1 ± 24.0EF) degree-days.

Therefore, at least two estimates are available for the duration of egg development from oviposition to hatch for each species. Mean estimates of 16 and 19 d for the eggs of S. vicinus are presented by Braman (1993) and Hayslip (1943), respectively. The shorter estimate is from a controlled temperature experiment, and may reflect the effect of higher constant temperature in the laboratory. The longer estimate is from the field, but represents an average including a wide range of incubation times over the course of the season.

Worsham and Reed (1912) do not give a mean duration for S. vicinus eggs, but report a range of 24–26 d, suggesting a 25 d average that shows far less variation than other estimates. An estimate of 21 d for the eggs of S. borellii is suggested by the data of both

Hayslip (1943) and Braman (1993). It is important to recognize that a significant amount of variation has been documented at both constant and variable temperatures for both species.

Date estimates for the timing of oviposition. Mean date estimates for each of the three percentages of peak and cumulative hatch differ by less than one week with most differing

105 by only one day (cf Tables 3 and 4). Therefore, using either peak or cumulative date estimates to calculate when oviposition occurred will give similar results. However, both these estimates are derived from field counts of C1 nymphs of unknown age. Therefore, we must make the assumption that half the nymphs were younger than the mean duration of the first instar, and half were older. Estimates for the date of 25, 50 and 75% cumulative oviposition are reported in Table 8. These estimates were made using the mean dates of C1 cumulative abundance for each species and subtracting the duration of the egg stage (19 d and 21 d for S. vicinus and S. borellii, respectively) and half the duration of the first instar (7 d and 11 d for S. vicinus and S. borellii, respectively) to calculate the estimates of oviposition date for each species. Similar calculations using the earliest date of first detection of C1 nymphs in the samples (3 June, JD 154, for both species) suggests that the date of first oviposition would be 8 and 2 May (JD 128 and 122) for S. vicinus and S. borellii, respectively. However, none of these estimates take the documented variability in the duration of either eggs or nymphs into account. Both temperature and moisture are likely to significantly modify the timing of oviposition, hatch and duration of the first instar.

Given that some nymphs probably hatched prior to the date of first detection, and that development may occur more slowly in the field, the date of first oviposition may occur earlier than indicated. Using first detection date, the longest egg incubation times reported, and half the longest duration reported for the first instar suggest that both species may begin oviposition in mid-April.

We can also use degree-day information to modify the date estimate for oviposition. Soil degree-day accumulations at the end of May range from approximately 11.1–16.7 (20–30 F)

106 degree-days per day. Using the mean degree-days calculated for the duration of incubation from Braman (1993), and dividing by degree-days accumulated in the field indicate a range of 17–25 and 21–32 d for the eggs of S. vicinus and S. borellii, respectively. Similar calculations for half the duration of the first instar yield a range of 7–11 and 11–17 d for S. vicinus and S. borellii, respectively. Subtracting the range calculated for duration of the egg stage and half of the first instar from our median date of hatch suggests median oviposition occurred 21 May - 2 June (JD 141–153) for S. vicinus and 23 May - 9 June (JD 143–160) for

S. borellii. These dates compare favorably with the dates calculated using date and duration in days alone. Similar calculations from the date of first detection indicate the date of first oviposition as 28 April - 10 May (JD 118–130) for S. vicinus and 15 April - 2 May (JD

105–122) for S. borellii. However, degree-day accumulations in April and early May can be substantially less than at the end of May, and egg-chamber depth can be substantially deeper than the depth at which our temperature measurements were made, especially in dry conditions. Degree-day accumulation will likely be slower at greater depths. Therefore, as with the estimates using date alone, it is better to take the conservative approach and assume that first oviposition can occur in early to mid-April.

Temperature and Degree-days. The most common method of quantifying the effect of temperature on insect development is to sum the heat required for completion of each stage or instar as degree-days. This is best accomplished where the upper and lower temperature thresholds of development have been established and development has been quantified in temperature controlled experiments. Where development has be quantified in this way, degree-days accumulations in the environment can be used to predict the occurrence of

107 damaging insect stages, as long as the temperature is monitored where it can be directly related to the development of the insect. Unfortunately, the temperature limits and degree- days required for mole cricket development have not been quantified in this manner, so we attempted an alternative method of following degree-days and development in the field. In our study we used equipment that calculated the accumulation of degree-days in the soil using a base temperature of 10EC (50EF) generally recommended for turfgrass insects

(Shetlar and Herms 1999). The upper limit of 43EC (110EF) is less meaningful for soil organisms, and was only ever encountered at the surface of the soil. Soil temperatures this high were never observed at the 10 cm recording depth in our study.

The degree-day concept is difficult to apply in the case of highly mobile organisms like mole crickets because the nymphs and adults are able to select their temperature environment through vertical migration in the soil column. Little is known about the subsurface movements of mole crickets, but in the laboratory nymphs of S. borellii have been reported at 14 cm (Tsedeke 1979) and S. vicinus nymphs can dig down to a depth of 38 cm (15 inches) in as few as 13 d (Villani et al. 2002). Adults tunnels have been reported as deep as 80 cm in the laboratory (Tsedeke 1979), 35 cm in the field (Hayslip 1943), and golf course personnel in NC have reported finding them over a meter deep in some situations.

Soil type, soil moisture, compaction, depth of the root zone and other factors influence depth and movement substantially, and these same factors can also affect the temperature profile of the soil. Construction of a degree-day model would be easier if data on preferred temperature or the details of their spatial and temporal movements in the soil column were known. Unfortunately, only circumstantial evidence of movement and behavior is available.

108 Degree-days as an explanatory variable. There are many reasons why degree-days did not prove to be the most viable explanatory variable (useful predictor) of mole cricket development. Measurements were only made at four sites per year, and only at a single location per site. This does not take into account the variation in soil moisture, compaction, soil type or turf coverage at the site. The accumulations were highly influenced by site selection, and only provide a rough estimate of soil degree-days. Unit malfunctions further decreased the data available for comparison with development. However, the most important factor is likely the behavior of the mole crickets themselves. Mole crickets are highly mobile insects capable of modifying their exposure to temperature and moisture conditions, primarily through vertical movement in the soil column. This allows them to position themselves in a microclimate favorable for survival and development. Differences in feeding preference could result in differences in behavior between the two species, and differences in preferred depth are indicated by studies on behavior (Tsedeke 1979) and tunnel architecture (Brandenburg et al. 2002, Villani et al. 2002). Theses factors complicate the development of a degree-day model predicting nymph abundance using 10 cm degree- day accumulations.

Moisture Environment of the Mole Cricket. Early in the research it became apparent that a model utilizing only soil degree-days was insufficient to explain differences in hatch date. Therefore, soil moisture was considered as another measurable variable that could modify the degree-day model and improve the prediction of hatch date. Previous experiments demonstrate there is a relationship between soil moisture and the timing of mole cricket oviposition in the laboratory (Hertl et al. 2001), and suggest that there may be

109 an effect on oviposition in the field. The results show that females prefer to oviposit in moist soil and delay oviposition under dry soil conditions. Soil moisture also affects other aspects of development and behavior. Variation in the depth of the egg chamber in response to soil moisture conditions has been reported, with the ovipositing females placing egg chambers at greater depths during dry conditions (Van Zwaluwenburg 1918, Hayslip 1943).

This behavior could affect duration of the egg stage as degree-day accumulations are likely to decrease with depth. The depth of the egg chamber might also effect the collection of nymphs during flush sampling. Moist soil condition at sampling allow greater penetration of the flush, but moist conditions earlier in the season would result in both the egg chamber and the nymphs being closer to the surface. Moisture may also have a direct effect on survival and movement of the nymphs, and dry soil conditions are not so easily avoided by vertical migration as high temperature.

We monitored soil moisture directly during the sampling season, however, oviposition took place much earlier, making these data less than useful for our purposes. Although most of our studies took place at irrigated sites, observations on turfgrass quality and vigor indicate that rainfall still had a significant influence on soil moisture levels. Therefore, we attempted to use rainfall data as an index of soil moisture. Rainfall increases soil moisture and could affect the timing of oviposition, incubation time, egg-hatch, and survival and duration of the first instar. Therefore, it is not surprising that rainfall was correlated with hatch date. The correlation between hatch and rainfall in May suggests that the primary effect was either on the timing of oviposition or duration of incubation. Correlation with

June rainfall may be due to other factors.

110 Rainfall as an explanatory variable. Like degree-days, rainfall was correlated with timing, but did not explain a high percentage of the difference observed in date among years or sites. Attempts to model hatch date and nymph development using both degree-days and rainfall were only partially successful, and may have been complicated by an interaction between temperature and moisture affecting behavior and development.

Irrigation is a necessity for maintenance of heathy turfgrass on golf courses and was a factor that provided both positive and negative benefits to the study. Without irrigation, dry soil conditions would have reduced sampling efficacy and the number of nymphs recovered.

Given that low nymph counts prevented the use of data from many sites, lower numbers would have further reduced the amount of usable data and our ability to characterize hatch.

Like rainfall, irrigation increases soil moisture and probably modifies the timing of oviposition and hatch. Our investigations took place at sites that are normally irrigated, therefore, we must accept irrigation as a physical factor modifying mole cricket development in the highly manipulated environment that such sites represent. Although this may obscure the natural relationship between mole cricket development and rainfall, the importance of rainfall is indicated by the fact that we were still able to find a relationship with timing in spite of the effects of irrigation.

CONCLUSION

The purpose of the study was to quantify hatch and model development to determine the best time to implement control. Estimates of hatch were quantified by date and soil degree- days for the first two size classes of nymphs for both species, and represent the first such estimates for this region. Although the relationship between size class and instar is unclear,

111 the two size classes examined represent the instars that are considered most susceptible to insecticides and are the targets for the most effective control. The date estimates for C1 and

C2 nymphs are similar, and can be considered as estimates for ‘treatable’ nymphs. These estimates were made directly from the actual abundance of nymphs observed, and our sample numbers represent the population of nymphs that were close enough to the surface to be contacted by the drench. Estimates for date of oviposition are subject to variation in the duration of the egg stage and first instar, both of which may be modified by environmental factors.

Our results indicate that abundance of the smaller mole cricket instars is better quantified by date than soil degree-day accumulations. Although both degree-days and rainfall explain some portion of the variability in date among years and sites, the amount of variation explained is not sufficient to produce a significant predictive model of timing. Correlation between these two independent variables further complicates analysis, however, the unknown physiological and behavioral responses to temperature and moisture probably have a greater effect. Species, nymph size, vertical movement, time of day, time of season, temperature, soil moisture, soil type, and turf cover probably all had some affect on sampling efficiency and results. Low numbers of nymphs at many sites limited the number of site years available for analysis and ability to develop a definitive model to predict development. However, date estimates for peak and cumulative abundance closely agree and provide useful guidelines for timing sampling and control. Information on degree-day accumulations and rainfall can be used to adjust these dates from year to year, and species composition at the site should also be taken into account. For example, managers at sites

112 with mixed populations, higher S. borellii populations, or sites that have experienced low rainfall in May and June should probably treat at the later end of the date or degree-day range. However, the decision to treat in all cases should be based on actually determining the presence and location of nymphs using an irritant flush. Our yield of nymphs from previously damaged areas indicates that mole cricket oviposition and hatch reoccur in the same locations from year to year. Therefore, mapping of visible damage in spring and fall should be used to define the areas where flushes should be performed and treatment may be needed.

Although soil degree-days and rainfall were both correlated with timing neither fully explain differences in date among years. Quantifying the exact nature of the relationship in the field will require information on soil moisture levels at the time of oviposition, and experiments to determine the effects of temperature and moisture on egg incubation time, hatch and nymph development. Additionally, the effects of temperature and moisture may vary substantially between stages and species.

Acknowledgments

The authors would like to extend their sincere thanks to the superintendents and management at the golf courses that participated in the study for making this project possible, and student research assistants Ian Winborne and Ed Karoly for help with the field sampling program. Special thanks are also extended to Dr. Cavell Brownie (Dept. of

Statistics, NC State University) for indispensable assistance with the statistical analysis of results. This work was funded in part by a grant from the United States Golf Association

Greens Section Research.

113 References Cited

Braman, S. K. 1993. Progeny production, number of instars, and duration of development of tawny and southern mole crickets (Orthoptera: Gryllotalpidae). J. Entomol. Sci.

28(4):327–330.

Brandenburg, R. L. and C. B. Williams. 1993. A complete guide to mole cricket management in North Carolina. NC Coop. Ext. Ser. ENT/ort - 101 8 pp.

Brandenburg, R. L., Y. Xia and A. S. Schoeman. 2002. Tunnel architecture of three species of mole crickets (Orthoptera: Gryllotalpidae). Fla. Entomol. 85(2): 383–385.

Cobb, P. P. and T. P. Mack. 1989. A rating system for evaluating tawny mole cricket,

Scapteriscus vicinus Scudder, damage (Orthoptera: Gryllotalpidae). J. Entomol. Sci.

24(1):142–144.

Forest, T. G. 1986. Oviposition and maternal investment in mole crickets (Orthoptera:

Gryllotalpidae): effects of season, size and senescence. Ann. Entomol. Soc. Am. 79(6):

918–924.

Forest, T. G. 1987. Insect size tactics and developmental strategies. Oecologia 73:

178–184.

114 Fowler, H. G. 1989. Natural microbial control of populations of mole crickets (Orthoptera:

Gryllotalpidae: Scapteriscus borellii): regulation of populations aggregated in time and space. Rev. Brasil. Biol. 49(4): 1039–1051.

Hayslip, N. C. 1943. Notes on biological studies of mole crickets at Plant City, Florida.

Fla. Entomol. 26(3): 33–46.

Hertl, P. T. and R. L. Brandenburg. 2002. Effect of soil moisture and time of year on mole cricket (Orthoptera: Gryllotalpidae) surface tunneling. Environ. Entomol. 31(3):

476–481.

Hertl, P. T., R. L. Brandenburg and C. B. Williams. 2003. Flight activity of

Scapteriscus vicinus Scudder and S. borellii Giglio-Tos (Orthoptera: Gryllotalpidae) in southeastern North Carolina. (Manuscript in Preparation).

Hudson, W. G. 1985. Ecology of the tawny mole cricket, Scapteriscus vicinus

(Orthoptera: Gryllotalpidae): Population estimation, spatial distribution, movement, and host relationships. Ph.D. dissertation, University of Florida.

Hudson, W. G. 1987. Variability in development of Scapteriscus acletus (Orthoptera:

Gryllotalpidae). Fla. Entomol. 70(3): 403–404.

115 Hudson, W. G. 1989. Field sampling and population estimation of the tawny mole cricket

(Orthoptera: Gryllotalpidae). Fla. Entomol. 72(2): 337–343.

Jessen, R. J. 1978. Statistical Survey Techniques. John Wiley and Sons. New York. 520 pp.

Kleyla, P. C. and G. Dodson. 1978. Calling behavior and spatial distribution of two species of mole crickets in the field. Ann. Entomol. Soc. Am. 71(4): 602–604.

Lawrence, K. O. 1982. A linear pitfall trap for mole crickets and other soil arthropods.

Fla. Entomol. 65(3): 376–377.

Matheny, E. L., Jr. 1981. Contrasting feeding habits of pest mole cricket species. J. Econ.

Entomol. 74(4): 444–445.

Matheny, E. L. and B. Stackhouse. 1980. Seasonal occurrence and life cycles data for S. acletus and S. vicinus, field-collected in Gainesville, Florida. Ann. Rep. No. 2, Mole

Cricket Research 79–80: 19–24.

SAS Institute. 2000. SAS/STAT user’s guide, version 8e. SAS Institute, Cary, NC.

116 Shetlar, D. J. and D. Herms. 1999. Insect and mite control on woody ornamentals and herbaceous perennials. Ohio State University Extension. Columbus OH. 73pp.

Short, D. E. and P. G. Koehler. 1977. Control of mole crickets in turf. Fla. Entomol.

60(2):147–148.

Short, D. E. and P. G. Koehler. 1979. A sampling technique for mole crickets and other pests in turfgrass and pasture. Fla. Entomol. 62(3):282–283.

Tsedeke, A. 1979. Plant material consumption and subterranean movements of mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus) as determined by radioisotopes techniques, with notes on materials for laboratory feeding. M.S. thesis. University of

Florida, Gainesville. 72 pp.

Ulagaraj, S. M. 1975. Mole crickets: ecology, behavior, and dispersal flight (Orthoptera:

Gryllotalpidae: Scapteriscus). Environ. Entomol. 4(2): 265–273.

Villani, M. G., L. L. Allee, L. Preston-Wilsey, N. Consolie, Y. Xia, R. L. Brandenburg.

2002. Use of radiography and tunnel castings for observing mole cricket (Orthoptera:

Gryllotalpidae) behavior in soil. Am. Entomol. 48(1): 42–50.

117 Walker, S. L. 1979. Population estimation, damage evaluation and behavioral studies on the mole crickets Scapteriscus vicinus and S. acletus (Orthoptera: Gryllotalpidae). M.S. thesis. University of Florida, Gainesville. 83 pp.

Walker, T. J. 1982. Sound traps for sampling mole cricket flights (Orthoptera:

Gryllotalpidae: Scapteriscus). Fla. Entomol. 65(1):105–110.

Walker, T. J. and D. A. Nickle. 1981. Introduction and spread of pest mole crickets:

Scapteriscus vicinus and S. acletus reexamined. Ann. Entomol. Soc. Am. 74(2):158–163.

Walker, T. J. and J. L. Nation. 1982. Sperm storage in mole crickets: fall mating fertilize spring eggs in Scpteriscus acletus. Fla. Entomol. 65(2): 283–285.

Williams, J. J. and L. N. Shaw. 1982. A soil corer for sampling mole crickets. Fla.

Entomol. 65(1): 192–194.

Worsham, E. L., and W. V. Reed. 1912. The mole cricket (Scapteriscus didactylus). Ga.

Exp. Sta. Bull. 101: 251–263.

Van Zwaluwenburg, R. G. 1918. The changa or West Indian mole cricket. Porto Rico

Agric. Exp. Stn. Bull. 23: 1–28.

118 Table 1. Pronotal midline lengths (mm) used to assign field collected Scapteriscus vicinus and S. borellii nymphs to size classes ( C1 - C7) equivalent to instar (adapted from Matheny and Stackhouse 1980).

Pronotal midline length (mm)

Size Class S. vicinus S. borellii

C1 1.4 – 1.6 1.6 – 1.9

C2 1.7 – 2.2 2.0 – 2.6

C3 2.3 – 3.0 2.7 – 3.9

C4 3.1 – 3.9 4.0 – 5.2

C5 4.0 – 5.2 5.3 – 6.4

C6 5.3 – 6.7 6.5 – 9.8

C7 6.8 – 9.9 C8 Adult winged winged

119 Table 2. Percentage species composition of two species of mole crickets from soapy water flush samples (n=10 in 1993–1996, n=12 in 1997) at two sites in North Carolina. FS = Fox

Squirrel, ST = Sea Trail

Site Year % S. borellii % S. vicinus Total Nymphs FS 1993 36.4 63.6 1031 FS 1994 28.6 71.4 1202 FS 1995 48.8 51.2 1372 FS 1996 88.3 11.7 695 FS 1997 87.4 12.6 864

ST 1993 37.7 62.3 326 ST 1994 16.1 83.9 1017 ST 1995 2.1 97.9 520 ST 1996 0.8 99.2 782 ST 1997 0.0 100.0 991

120 Table 3. Julian date estimates for three percentages of peak mean count for two size classes of S. vicinus and S. borellii in southeastern North Carolina (1993–1997).

Species and Julian Date Estimates for % of Peak Nymph Abundance

Instar 25 % 50 % 75 % Instar 1 = C1 S. vicinus 171 176 178 S. borellii 182 191 201 Instar 2 = C2 S. vicinus 170 176 183 S. borellii 190 200 207

Table 3a. Calendar date estimates for three percentiles of peak mean count for two size classes of S. vicinus and S. borellii in southeastern North Carolina (1993–1997).

Species and Calendar Date Estimates for % of Peak Nymph Abundance

Instar 25 % 50 % 75 % Instar 1 = C1 S. vicinus 20 June 25 June 27 June S. borellii 1 July 10 July 20 July Instar 2 = C2 S. vicinus 19 June 25 June 2 July S. borellii 9 July 19 July 26 July

121 Table 4. Julian date estimates (± SE) for three percentages of cumulative abundance for two size classes of S. vicinus and S. borellii in southeastern North Carolina (1993–1997).

Species and Julian Date Estimates for % of Cumulative Nymph Abundance

Instar 25 % 50 % 75 % Instar 1 = C1 S. vicinus 170.9 ± 2.0 177.4 ± 1.9 183.9 ± 2.0 S. borellii 185.7 ± 2.2 192.3 ± 2.1 198.9 ± 2.2 Instar 2 = C2 S. vicinus 174.9 ± 1.6 181.9 ± 1.5 188.8 ± 1.7 S. borellii 192.1 ± 1.9 198.2 ± 1.7 204.3 ± 1.9 Data are LS Mean estimates of date among years and sites. N for C1 = 13 and 14, N for C2

= 15 and 12, for S. vicinus and S. borellii, respectively.

Table 4a. Calendar date estimates (± SE) for three percentages of cumulative abundance for two size classes of S. vicinus and S. borellii in southeastern North Carolina (1993–1997).

Species and Calendar Date Estimates for % of Cumulative Nymph Abundance

Instar 25 % 50 % 75 % Instar 1 = C1 S. vicinus 20 June ± 2.0 26 June ± 1.9 3 July ± 2.0 S. borellii 5 July ± 2.2 11 July ± 2.1 19 July ± 2.2 Instar 2 = C2 S. vicinus 24 June ± 1.6 1 July ± 1.5 8 July ± 1.7 S. borellii 11 July ± 1.9 17 July ± 1.7 23 July ± 1.9 Data are LS Mean ( ± SE) estimates of date among years and sites. N for C1 = 13 and 14, N for C2 = 15 and 12, for S. vicinus and S. borellii, respectively.

122 Table 5. Summary of mean soil degree-day accumulations (DDEC ± SD) among sites and years at three percentages of cumulative abundance for two size classes of S. vicinus and S. borellii in southeastern North Carolina (1993–1997; n = 11 for S. vicinus; n = 11 and 10 site- years for C1 and C2 S. borellii, respectively).

Species and Degree-Day Accumulations for Percentage of Nymph Abundance

Instar 25 % 50 % 75 % Instar 1 = C1 S. vicinus 1108.6 ± 62.2 1218.4 ± 70.8 1332.5 ± 100.0 S. borellii 1363.9 ±146.3 1483.5 ± 143.4 1606.6 ± 140.1 Instar 2 = C2 S. vicinus 1179.4 ± 105.8 1295.2 ± 126.7 1422.7 ± 153.1 S. borellii 1506.3 ± 156.9 1627.2 ± 145.8 1752.1 ± 162.3

123 3 5 11 11 2 3 2 2 3 1 1 202.8 1693.3 188.8 ± 4.6191.0 ± 1.3 1359.8 ± 121.7 177.9 ± 2.8180.8 ± 1.2 1556.7 1302.2 ± 33.0 181.2 ± 4.3 1291.7 ± 26.4 183.9 ± 5.6 1270.8 ± 31.1 1332.5 ± 100.0 193.9 ± 0.7195.8 ± 9.3 1520.6 ± 69.2 1592.6 ± 63.7 198.5 ± 8.2 1606.6 ± 140.1 200.6 ± 10.7 1632.1 ± 198.2 5 11 3 2 3 2 2 3 11 1 1 C) from 1 January for three percentages of cumulative C) from 1997), N = sites for date analysis, DD superscript no. E – 196.2 1558.9 182.5 ± 1.3 1233.9 ± 55.6 182.6 ± 0.7173.0 ± 3.7 1373.9 1220.0 ± 52.7 175.5 ± 1.7 1196.8 ± 38.9 173.5 ± 4.4 1148.3 ± 45.6 187.3 ± 2.6 1404.4 ± 22.8 190.2 ± 7.3 1484.1 ± 50.7 192.1 ± 8.1 1483.5 ± 143.4 194.0 ± 11.3 1499.7 ± 212.4 177.6 ± 4.8 1218.4 ± 70.8 5 11 3 2 3 2 3 2 11 1 1 in southeastern NC, (1993 25% Hatch25% Hatch 50% Hatch 75% S. borellii and N Date DD ± SD Date DD ± SD Date DD ± SD S. vicinus 14 185.6 ± 7.9 1363.9 ± 146.3 13 171.2 ± 4.5 1108.6 ± 62.2 Comparison of Julian date and degree-days (DD, base 10 Comparison 19931994 31995 176.2 ± 2.5 21996 174.1 ± 2.8 1122.6 ± 46.4 21997 168.2 ± 4.5 4 1195.6 170.2 ± 2.5 1147.2 ± 51.1 2 165.9 ± 4.6 1095.2 ± 53.7 1993 1025.6 ± 23.6 1994 51995 187.4 ± 12.3 2 1371.8 ± 219.9 1996 186.3 ± 7.1 21997 180.6 ± 4.5 4 184.7 ± 5.4 1297.2 ± 1.6 1 --- 1364.8 ± 39.4 189.7 193.9 ± 5.7 1455.0 --- 201.5 ± 4.2 --- Year Mean Mean S. vicinus S. borellii DD correlation analysis. Table 6. abundance for 124 1997), N = sites for date. – C1 nymphs in southeastern NC, (1993 C1 nymphs C) from 1 January to 30 May and monthly rainfall accumulations used for rainfall accumulations 1 January to 30 May and monthly C) from E S. vicinus Julian Date of Hatch ± SD CM of Rainfall ± SD N 25% 50% 75% DD ± SD May June 11 171.2 ± 4.5 177.6 ± 4.8 183.9 ± 5.6 778.9 ± 66.1 5.1 ± 2.0 14.7 ± 7.9 Degree-day accumulations (DD, base 10 Degree-day accumulations 19931994 31995 1 176.0 ± 2.71996 2 182.7 ± 1.5 172.01997 3 168.0 ± 4.2 189.0 ± 4.6 2 170.3 ± 3.1 173.0 ± 4.2 166.0 ± 4.2 688.3 ± 11.5 182.0 176.0 ± 2.0 178.0 ± 2.8 173.5 ± 5.0 4.1 ± 0.5 181.3 ± 1.5 847.8 ± 8.7 181.0 ± 4.2 192.0 771.7 ± 0.0 7.9 ± 2.0 5.6 ± 2.8 835.6 ± 27.1 4.8 ± 0.0 822.2 3.8 ± 2.5 28.5 ± 4.1 17.0 ± 3.6 15.0 ± 5.6 3.3 13.7 Year Mean modeling date of cumulative abundance for date of cumulative modeling Table 7.

125 Table 8. Comparison of effects for 30 May degree-day accumulations and June rainfall as explanatory variables for date of 50% cumulative abundance of S. vicinus C1 nymphs. (df = 1 for both variables).

Parameter Estimate ± SE t value P value (R2 = 0.56) Intercept 220.1 ± 12.7 17.36 < 0.0001 30 May DD -0.03 ± 0.0 -3.38 0.0082 (R2 = 0.52) Intercept 184.0 ± 2.1 88.45 < 0.0001 June Rain -1.11 ± 0.3 -3.47 0.0053 (R2 = 0.64) Intercept 201.8 ± 17.9 11.30 < 0.0001 June Rain -0.71 ± 0.51 -1.39 0.2031 30 May DD -0.01 ± 0.0 -1.00 0.3467

Table 9. Calendar date estimates for three percentages of cumulative oviposition for S. vicinus and S. borellii in southeastern North Carolina (1993–1997).

Calendar Date Estimates for % of Cumulative Oviposition Species 25 % 50 % 75 % S. vicinus 25 May 31 May 7 June S. borellii 3 June 9 June 16 June Estimates for duration of egg stage and 0.5 first instar from Hayslip (1943), Braman (1993).

126

3

S. vicinus

2 ++

S. borellii

1 + 127

0

3-Jul 8-Jul 8-Jun 2-Aug 13-Jul 18-Jul 23-Jul 28-Jul 13-Jun 18-Jun 23-Jun 28-Jun CALENDAR DATE

Figure 1. Comparison of timing between peak mean counts of C1 nymphs of S. vicinus and

S. borellii. Data are mean soapy water flush counts taken on golf courses in southeastern

North Carolina (1993–1997). 3

S. vicinus

2 +

S. borellii

1 + 128

0

3-Jul 8-Jul 8-Jun 2-Aug 13-Jul 18-Jul 23-Jul 28-Jul 13-Jun 18-Jun 23-Jun 28-Jun CALENDAR DATE

Figure 2. Comparison of timing between peak mean counts of C2 nymphs of

S. vicinus and S. borellii. Data are mean soapy water flush counts taken on golf

courses in southeastern North Carolina (1993–1997). -0.9

-0.8 25% -0.7

-0.6

-0.5 50%

-0.4

-0.3 SIG CORR (P < 0.05) -0.2 129 NON SIG CORR

PEARSON CORR. COEFFICIENT -0.1

0 9 July 9 June 19 July 29 July 10 May 20 May 30 May 10 April 20 April 30 April 19 June 29 June DATE

Figure 3. Trend in correlation between date of 25 and 50% cumulative abundance of Class 1

S. vicinus nymphs and soil degree-day accumulations at twelve dates from golf courses in southeastern

North Carolina (n = 11 site-years, 1993–1997). Chapter IV

The Effect of Soil Moisture and Time of Year on Mole Cricket (Orthoptera:

Gryllotalpidae) Surface Tunneling

Peter T. Hertl and Rick L. Brandenburg

______

ABSTRACT The damage caused by two species of introduced mole cricket pests

(Scapteriscus vicinus Scudder and Scapteriscus borellii Giglio-Tos) was studied during

1995, 1996, and 1997 in Brunswick County, North Carolina. Surface tunneling activity in bermudagrass was quantified weekly starting in late July or early August using a modification of the damage grid evaluation method of Cobb and Mack (1989). Soil moisture was monitored in three depth ranges (0–10.2, 10.2–20.3, and 20.3–30.5 cm) and percentage soil moisture based on dry soil weight was determined gravimetrically. Most sites had a higher percentage of S. borellii than S. vicinus, with the percentage of S. borellii ranging from 42–95%. The percentage soil moisture ranged from 1.2–24.9, 1.3–19.3, and

1.3–20.4% at 0–10.2, 10.2–20.3, and 20.3–30.5 cm, respectively. Moisture percentages in the 10.2–20.3 and 20.3–30.5 cm ranges were not significantly different. Mean percentage soil moisture in the 0–10.2 cm range was significantly greater than for the 10.2–20.3 and

20.3–30.5 cm ranges combined. Average damage ratings (0–9) increased linearly with

Julian date, but due to differences in damage levels among the years, three separate linear

130 equations were used to describe the relationship. Mean damage ratings increased by one rating point (11%) every 2–3 weeks. A significant nonlinear relationship was found between percentage soil moisture and mean damage ratings. Management implications of the findings are discussed.

Key Words: Mole crickets, Scapteriscus, surface activity, behavior, damage, soil moisture

______

Mole crickets are economically important soil-dwelling pests that damage turfgrass and a variety of other crops throughout the southeastern United States. As is true of many subterranean insects, their behavior is difficult to study because direct observation of their activities is not feasible. One of the most obvious and observable aspects of mole cricket behavior is their tunneling activity at the soil surface. The disturbed soil resulting from this digging is often the only evidence of their presence in the habitat. Although the most serious damage to turfgrass occurs when the insects actually feed on the grass itself, mole cricket surface tunneling in managed turf is usually considered as damage, even when the mounding and tunneling are not associated with feeding on the roots and stems. The disturbed soil can disrupt play on golf greens and tees, cause unsure footing on playing fields, damage mowing equipment, and ruin the appearance of what would otherwise be considered well-managed turf. For the purposes of this paper, the terms surface tunneling, surface activity and damage are used synonymously.

There are two species of introduced mole cricket pests found in North Carolina turfgrass.

The tawny mole cricket, Scapteriscus vicinus Scudder is primarily a phytophagous species,

131 feeding on the roots and stems of the grass. The southern mole cricket, Scapteriscus borellii

Giglio-Tos is more of a predaceous species, feeding mainly on other soil arthropods, but also feeding occasionally on the grass (Matheny 1981). The surface tunneling of both species is quite variable throughout the season, and is dependent on a variety of factors including soil temperature, stage, and the size of the insect. There may also be differences in surface tunneling between the two species. Hayslip (1943) studied both S. vicinus and S. borellii [acletus] in Florida and noted that S. borellii caused a greater proportion of the surface burrowing than S. vicinus, but the latter species was more responsible for direct feeding injury to the plants. Differences in tunneling between the sexes have also been investigated. Walker (1979) performed a small study (five males, five females) on S. borellii surface tunneling activity in buckets. No differences were found in tunneling activity between the sexes, however, there was high variation in tunneling among individuals, and mean tunneling activity tended to increase with time.

Another factor that may affect surface tunneling is soil moisture. The hypothesis that an increase in soil moisture stimulates mole cricket surface activity has often been suggested, but has never been quantified, or substantiated by experiment. Barrett (1902) commented extensively on the natural history of Scapteriscus didactylus (Latreille) [as vicinus] in Puerto

Rico, stating that the crickets avoided dry soil conditions and remained at a depth of several inches unless the surface of the soil was moist. He also reported that the greatest amount of damage occurs during moist soil conditions, and suggested killing mole crickets at the surface after rain as a means of control. Similarly, Worsham and Reed (1912) reported that

S. vicinus was more active at the surface during moist conditions in Georgia. Hayslip

132 (1943) also reported that most mole cricket surface feeding occurs when the soil was warm and moist, and that in dry fields, they preferred to burrow and feed around transplants that had been freshly watered.

Our field observations in North Carolina also indicate that soil moisture dramatically influences surface tunneling. Mole crickets are apparently more active during times of high soil moisture and much less active during times of low soil moisture. Further evidence for a relationship between soil moisture and surface tunneling is the fact that mole cricket damage is often more serious near irrigation heads than in the surrounding turf. Although the effect of soil moisture on mole cricket tunneling has not been addressed experimentally, Hudson

(1985) studied the surface movements of S. vicinus in bahiagrass pastures using arenas and linear pitfall traps. He found no effect on surface movement within soil ranging between

9.5–15.1% moisture in the sampling arenas and 10.0–20.8% moisture in the field pitfall study, but did note increased activity following rainfall if the soil had been dry.

Due to the apparent relationship between mole cricket surface activity and soil moisture, it has become a standard practice to recommend the irrigation of infested areas prior to applying insecticides for the control of these insects. The increased moisture levels are believed to stimulate the surface activity of the pest, increasing their exposure to the insecticidal agents and improving the efficacy of the treatment. Because of the important management implications of the relationship between soil moisture and surface activity, it is important to fully understand the insect’s behavior in regard to this factor. This study was performed to elucidate the relationship between the surface tunneling activity of mole crickets and soil moisture in the turfgrass environment.

133 Materials and Methods

Study Site. This study was conducted in 1995, 1996, and 1997 at The Lakes Country

Club (formerly the Fox Squirrel Country Club, 34E 02' N, 78E 03' W) in Brunswick County,

North Carolina. The nine bermudagrass (Tifgreen 328) fairways used in the study have a long history of serious mole cricket infestation and are composed mainly of Kureb fine sand, with one site each in Mandarin and Leon fine sand. Each year separate untreated portions of infested fairways were selected as test sites and four randomly chosen plots within each site were selected for use in the study. These plots consisted of either untreated test plots (6 m X

6 m or larger) in field insecticide efficacy trials, or other untreated fairway areas that could be repeatedly sampled in a similar fashion. All attempts were made to locate each study site in an area that was uniform in both level of infestation and soil type. Eight sites were sampled in 1995, and six sites were sampled each year in 1996 and 1997.

Species Composition, Surface Activity and Soil Moisture. Species composition on each fairway was determined by soapy water flush sampling (Short and Koehler 1979) conducted on one date in mid to late July, after the majority of the eggs of both species had hatched.

Voucher specimens of both species were placed in the North Carolina State University insect collection. Surface activity was quantified using a modification of the standard damage grid evaluation method of Cobb and Mack (1989). The original method was developed using a 0.6 m2 grid divided into nine subsections, however, a 1.0 m2 PVC grid was used in this study. Damage ratings were made by randomly placing the grid in the area selected for evaluation. Fresh damage was identified both visually and by touch, and a numeric damage rating (0–9) was given based on the occurrence of fresh damage (surface

134 tunneling of any kind) in the nine subsections of the grid (with 0 = no damage, 9 = damage in all nine subsections). Damage ratings were made weekly starting in late July or early

August, beginning within 1–2 wk after the surface activity from the current year’s nymph population became visible enough to produce non-zero ratings in fairway-cut turf. Five ratings were made in each of the four replicates per site on each sampling date and soil moisture samples were taken at the time of each evaluation. Soil moisture was monitored at three depths by taking two random soil cores from each of the four plots evaluated for surface activity at each site. The top cm of grass and thatch was removed and each core was separated into portions representing 0–10.2, 10.2–20.3 and 20.3–30.5 cm (0–4, 4–8, and

8–12 inches, respectively) in depth in the soil profile. The sample of cores from each depth range were combined as a composite sample, sealed in plastic bags, and chilled to help prevent moisture loss. The samples were returned to the lab and percentage soil moisture

(based on dry soil weight) was determined by the gravimetric method described by Kramer

(1969). This method estimates the weight of the water in the soil by calculating the difference in weight of the sample before and after oven drying. To enable the comparison of the percentage moisture data reported here with those in other studies, a standard matric potential curve for Kureb fine sand from the study site was developed using the standard pressure plate extractor technique following the methods described by Kramer (1969).

Data Analysis. Soil moisture data were analyzed using repeated measures analysis with depth as a repeated measures factor. Orthogonal contrasts were used to compare the percentage moisture at 10.2–20.3 cm with that at 20.3–30.5 cm; as well as the percentage moisture at 0–10.2 cm with the mean of the percentage moisture in the 10.2–20.3 and

135 20.3–30.5 cm ranges combined. Mean damage ratings and mean percentage soil moisture within each depth range were analyzed using ANOVA, and tested for significant differences among years (averaged over dates within years) using the least significant difference (LSD) means separation test. The relationship between damage and Julian date was examined using a modified analysis of covariance (PROC GLM) using a model including terms for year, block within year, date linear and date quadratic, date linear by year, and date quadratic by year. The relationship between mean surface tunneling activity, as quantified by grid ratings, and mean percentage soil moisture was examined using a modified analysis of covariance (PROC GLM) using a model including terms for year, block within year, date, and various combinations of linear and quadratic terms for percentage moisture (SAS

Institute 1999).

Results

Species Composition. The results of soapy water flush sampling indicate that most sites had a higher percentage of S. borellii than S vicinus. The percentage of S. borellii at the sites ranged from 54–90% in 1995, 42–95% in 1996 and 83–95% in 1997. Excluding one site (42%), the percentage of S. borellii in 1996 was 78–95%. Additionally, the results of replicated weekly flush sampling on 10 dates (end of May through July) in a concurrent study on mole cricket development indicate that the overall percentage of S. borellii on the course was 48.8% in 1995, 88.3% in 1996 and 88.4% in 1997. Numbers of S. borellii recovered in the samples were similar among the years, indicating substantial lower S. vicinus populations in 1996 and 1997 (unpublished data).

Soil Moisture. Mean percentage soil moisture for each of the three depths is reported by

136 year in Table 1. The percentage moisture at 0–10.2 cm ranged from 2.0–23.7% in 1995,

5.9–24.9% in 1996 and 1.2–11.2% in 1997. The percentage moisture at 10.2–20.3 cm ranged from 1.3–17.1% in 1995, 2.9–19.3% in 1996 and 1.3–7.4% in 1997. The percentage moisture at 20.3–30.5 cm ranged from 1.5–17.0% in 1995, 3.9–20.4 in 1996 and 1.3–8.3% in 1997. The overall mean percentage moisture for the three years of study was 9.4 ± 0.5,

6.3 ± 0.3 and 6.4 ± 0.4% at 0–10.2, 10.2–20.3 and 20.3–30.5 cm, respectively.

There were significant differences (F = 250.35; df = 2, 146; P < 0.0001) in the percentage soil moisture at different depths. Orthogonal contrasts show that the percentage moisture at 10.2–20.3 cm and 20.3–30.5 cm were not significantly different (F = 1.48; df =

1, 73; P = 0.2280), and could be combined. The mean percentage moisture at 0–10.2 cm was significantly greater than the mean percentage moisture in the 10.2–20.3 and 20.3–30.5 cm ranges combined (F = 378.5; df = 1, 73; P < 0.0001). There were no significant differences among the years with regard to mean percentage moisture at the 20.3–30.5 cm level. However, mean percentage moisture at both the 0–10.2 and 10.2–20.3 cm levels were significantly greater in 1995 than in 1997 (t = 2.11; df = 2, 17; P 0.05). Mean percentage moisture in 1996 at both the 0–10.2 and 10.2–20.3 cm levels were not significantly different from those observed in either 1995 or 1997 (t = 2.11; df = 2, 17; P > 0.05).

Percentage soil moisture data reported here is most meaningfully compared with those reported in other studies using matric potential (-KPa). A graph relating percentage moisture and matric potential for Kureb fine sand from our study site is presented in Fig. 1.

This graph shows that between 4–23% (-10 to 0 KPa, respectively), moisture is readily available to soil organisms, but at moisture levels less than 4% (-50 to -1500 KPa) moisture

137 is biologically unavailable in this soil. Kureb, Mandarin and Leon fine sand are similar in surface texture and the matric potential relationship will be comparable on a percentage basis.

Surface Activity and Julian Date. Mean damage ratings for each year are reported in

Table 1. Mean damage ratings were found to increase linearly with Julian date (Fig. 2).

Due to significant differences (F = 4.3; df = 2, 17; P < 0.05) in the level of damage during the three years of the study, a single linear equation could not be used to model mean damage ratings. The three linear equations describing the relationship between mean damage ratings and Julian date are: Y = -11.18 + 0.0752X (for 1995, r2 =0.59), Y = -12.85 +

0.0749X (for 1996, r2 =0.34) and Y = -8.38 + 0.0526X (for 1997, r2 =0.23) [Y = a + bX, where Y = average damage rating, a = intercept, b = slope, and X = Julian date]. The slopes of the lines indicate that the rate of increase in damage is approximately the same in 1995 and 1996, but slightly less in 1997. Mean damage ratings increased by one rating point

(11%) every 2–3 wk (13.3, 13.4 and 19.0 d for 1995, 1996 and 1997, respectively).

Surface Activity and Soil Moisture. A significant relationship was found between soil moisture and surface tunneling activity. Mean damage ratings were related to mean percentage moisture using two separate analyses, one of which included both a linear and quadratic term for mean percentage moisture in the 0–10.2 cm samples, and the second which included linear and quadratic terms for the mean of the combined percentage moisture in the 10.2–20.3 and 20.3–30.5 cm samples (Table 2). Due to differences in damage among the years, a highly significant block effect, and the significant relationship between mean damage ratings and Julian date, both analyses fitted a different quadratic for

138 each date and year. The analysis including terms for soil moisture from 0–10.2 cm is more significant with regard to the F and P-values for the quadratic moisture term, and is the more appropriate model for addressing the relationship with surface activity.

Discussion

Although the relationship between soil moisture and mole cricket surface tunneling has often been alluded to in the literature, this is the first study verifying these observations with a replicated field test. Our results show that surface tunneling increased significantly with increases in the percentage soil moisture in the upper soil layer, and that the relationship is nonlinear. Given the obvious relationship of moisture among the levels sampled, it is not surprising that moisture in the deeper samples could also be correlated with tunneling activity. The study also demonstrates that damage increases linearly with time of year

(Julian date). Although the rate of increase in damage ratings is similar, the level of damage can be substantially different from year to year. The differences observed in damage levels may have been due to several factors, including the differences noted in population composition and mean soil moisture among the years.

The damage grid rating system developed by Cobb and Mack (1989) is the only currently accepted method for quantifying mole cricket surface tunneling activity. It has become the standard technique used to evaluate mole cricket insecticide efficacy trials and provided a quick and easy method to quantify surface tunneling in our study. Although the rating system was developed for evaluating damage caused by S. vicinus, it has been widely used to evaluate the damage caused by S. borellii, as well as mixed populations of the two species. Unfortunately, it is virtually impossible to differentiate the damage caused by the

139 two species, and was not attempted in this study. The rating technique does have some limitations. It can only be used when the mole cricket nymphs are large enough to produce visible evidence of their surface activity, and can not accurately estimate population size when enough nymphs are present to consistently cause damage in all nine subsections of the grid (Cobb and Mack 1989).

The damage grid rating system has been widely adopted because of the difficulty of making mole cricket population estimates by other means. Cobb and Mack (1989) found that S. vicinus nymph counts from soapy water flush samples (Short and Koehler 1979) were linearly related to ratings by the equation Y= 0.57X - 0.78 (where Y = number of nymphs per

0.6 m2, X= damage rating). Both the mixed species composition of the populations, and the larger grid size used, made it impossible to use this equation to estimate population size in our study.

The need for a quadratic term for moisture in the analysis directly addresses the nonlinear nature of the relationship between moisture and surface activity. Many aspects of insect behavior, reproduction and physiology are not well-described by linear functions. This is especially true of moisture relationships among edaphic insects because they have adapted to survive and function over a wide range of moisture conditions, and are only sensitive to moisture extremes. This type of relationship was observed in the ovipositional response of

Popillia japonica Newman (Régnière et al. 1979), where oviposition in sandy soil did not take place below 3% moisture, was unimpeded at mid-range moisture levels, and was reduced at moisture levels approaching saturation. A similar relationship was shown for oviposition preference in Cerotoma trifurcata (Foster) (Marrone and Stinner 1983a), and

140 between egg-hatch and soil moisture in C. trifurcata (Marrone and Stinner 1983b). In these studies few eggs were laid, hatched or survived in loamy sand at moisture levels near or below the wilting point of plants, or in saturated soil. Egg-hatch was uniformly high between these extremes. Hertl et al. (2001) demonstrated that adult mole crickets suffer high mortality when confined in Kureb fine sand at 2% moisture, however, no differences in mortality were noted at soil moisture levels between 4 and 12%. The substantial differences in adult survival between 2 and 4% moisture, and the matric potential data for this soil suggests a threshold moisture level of approximately 3%. Data presented here show that mole crickets can be subjected to an extreme range of moisture near the soil surface.

However, our results indicate that mole crickets avoid dry conditions by limiting their activity at the surface during periods of low moisture, and resume their surface activities when acceptable moisture levels prevail.

The study related surface tunneling with existing differences in soil moisture due to differences in rainfall, soil moisture retention, and irrigation schedules. Although we did not utilize irrigation treatments to show differences in surface activity, irrigation is the primary means of increasing soil moisture in managed turfgrass. The study demonstrated that moisture does affect the surface activity of large nymphs capable of producing visible, rateable damage in late summer. However, most insecticide applications for mole cricket control are timed to coincide with peak hatch, which usually occurs in late June or early July in NC (unpublished data). During that period the population is composed primarily of very small nymphs, and damage is not evident. Although we cannot directly relate our findings to the smaller instars, it is likely that their behavioral response to moisture is similar.

141 Therefore, we agree with the practice of recommending irrigation to stimulate surface activity prior to applying insecticides for the control of mole crickets.

The study found a significant positive relationship between soil moisture and mole cricket surface activity. Previous studies have found significant relationships between soil moisture and mole cricket calling behavior (Ulagaraj 1976), egg chamber depth (Van

Zwaluwenburg 1918, Worsham and Reed 1912, Hayslip 1943) and oviposition (Hertl et al.

2001). Rainfall is another moisture-related factor affecting mole cricket behavior.

However, the evolutionary context of the response is the resulting increase in soil moisture levels, rather than the event itself. Rainfall has been cited as stimulating adult flights

(Hayslip 1943, Ulagaraj 1975, Walker 1982), and Hudson (1985) noted increased surface movement of nymphs following rainfall if the soil had been dry. Surface tunneling activity can be added to the list of mole cricket behaviors influenced by soil moisture, and we now have an empirical basis for considering and using soil moisture in the management of these pests.

Acknowledgments

The authors would like to extend their sincere thanks to superintendents Rick Vigland and

Chuck Baldwin of the Fox Squirrel Country Club, and student research assistants Ian

Winborne and Ed Karoly. Special thanks are also extended to Dr. Cavell Brownie (Dept. of

Statistics, NC State University) for indispensable assistance with the statistical analysis of results. This work was funded in part by a grant from the United States Golf Association

Greens Section Research.

142 References Cited

Barrett, O. W. 1902. The changa, or mole cricket (Scapteriscus didactylus Latr.) in Porto

Rico. P. R. Agric. Stn. Bull. 2: 1-19.

Cobb, P. T., and T. P. Mack. 1989. A rating system for evaluating tawny mole cricket,

Scapteriscus vicinus Scudder, damage (Orthoptera: Gryllotalpidae). J. Entomol. Sci. 24:

142-144.

Hayslip, N. C. 1943. Notes on biological studies of mole crickets at Plant City, Florida.

Fla. Entomol. 26: 33-46.

Hertl, P. T., R. L. Brandenburg, and M. E. Barbercheck. 2001. The effect of soil moisture on ovipositional behavior in the southern mole cricket Scapteriscus borellii Giglio-

Tos (Orthoptera: Gryllotalpidae). Environ. Entomol. 30: 466-473.

Hudson, W. G. 1995. Ecology of the tawny mole cricket, Scapteriscus vicinus (Orthoptera:

Gryllotalpidae): Population estimation, spatial distribution, movement, and host relationships. Ph.D. dissertation, University of Florida.

Kramer, P. J. 1969. Plant and soil water relationships: a modern synthesis. McGraw-Hill,

New York.

143 Marrone, P. G., and R. E. Stinner. 1983a. Effects of soil moisture and texture on oviposition preference of the bean leaf beetle, Cerotoma trifurcata (Foster) (Coleoptera:

Coccinellidae). Environ. Entomol. 12: 426-428.

Marrone, P. G., and R. E. Stinner. 1983b. Effects of soil physical factors on egg survival of the bean leaf beetle, Cerotoma trifurcata (Foster) (Coleoptera: Chrysomelidae). Environ.

Entomol. 12: 673-679.

Matheny, E. L., Jr. 1981. Contrasting feeding habits of pest mole cricket species. J. Econ.

Entomol. 74: 444-445.

Régnière, J., R. L. Rabb, and R. E. Stinner. 1981. Popillia japonica: Effects of soil moisture and texture on survival and development of eggs and first instar grubs. Environ.

Entomol. 10: 654-660.

SAS Institute. 1999. SAS/STAT users guide, version 8.0. SAS Institute, Cary, NC.

Short, D. E. and P. G. Koehler. 1979. A sampling technique for mole crickets and other pests in turfgrass and pasture. Fla. Entomol. 62: 282-283.

Ulagaraj, S. M. 1975. Mole crickets: ecology, behavior, and dispersal flight (Orthoptera:

Gryllotalpidae: Scapteriscus). Environ. Entomol. 4: 265-273.

144 Ulagaraj, S. M. 1976. Sound production in mole crickets (Orthoptera: Gryllotalpidae:

Scapteriscus). Ann. Entomol. Soc. Am. 69: 299-306.

Van Zwaluwenburg, R. G. 1918. The changa or West Indian mole cricket. Porto Rico

Agric. Exp. Stn. Bull. 23: 1-28.

Walker, S. L. 1979. Population estimation, damage evaluation and behavioral studies on the mole crickets Scapteriscus acletus and S. vicinus (Orthoptera: Gryllotalpidae). M.S.

Thesis, University of Florida, Gainesville.

Walker, T. J. 1982. Sound traps for sampling mole cricket flights (Orthoptera:

Gryllotalpidae: Scapteriscus). Fla. Entomol. 65: 105-109.

Worsham, E. L., and W. V. Reed. 1912. The mole cricket. Georgia Exp. Stn. Bull. 101:

251-263.

145 Table 1. Mean mole cricket damage ratings (0–9) and mean percentage soil moisture from bermudagrass fairways in NC (1995, 1996 and 1997)

Mean Damage Mean Percentage Moisture ± SE

YEAR N Ratings ± SE 0-10.2 cm 10.2-20.3 cm 20.3-30.5 cm 1995 51 6.0 ± 0.4a 11.0 ± 1.4a 7.4 ± 1.0a 7.4 ± 1.1a 1996 37 4.6 ± 0.5b 10.3 ± 1.9ab 6.7 ± 1.3ab 7.0 ± 1.4a 1997 26 4.1 ± 0.8b 5.2 ± 3.2b 3.7 ± 2.1b 3.7 ± 2.3a Means in the same column with the same letter are not significantly different (LSD, P0.05)

146 Table 2. Modified analysis of covariance of the relationship between mean damage ratings (0–9) and mean percentage soil moisture (%SM) at 0–10.2 cm (r2 = 0.96); and in the 10.2–20.3 and 20.3–30.5 cm ranges, combined (r2 = 0.95)

0–10.2 cm 10.2–30.5 cm

Variable df F-value P-value F-value P-value Year 2 8.09 0.0009 8.57 0.0006 Block (Year) 17 15.37 <0.0001 12.56 <0.0001 Date (Year) 18 8.59 <0.0001 10.20 <0.0001 %SM 1 3.73 0.0590 3.47 0.0684 %SM * %SM * Date (Year) 21 2.64 0.0024 2.03 0.0204

147 1600 1400 1200 1000 800 600 400 200 MATRIC POTENTIAL (- KPa) 0 0 2 4 6 8 1012141618202224 % SOIL MOISTURE

Figure 1. Relationship between soil moisture and matric potential

(-KPa) for Kureb fine sand from the study site.

148

9

8 1995

7

6 1996

5 1997

4

3 1995 2 1996 1997 1 MEAN DAMAGE RATING MEAN DAMAGE 0 7 SEP 7 OCT 8 AUG 19 JUL 29 JUL 17 SEP 27 SEP 17 OCT 18 AUG 28 AUG DATE

Figure 2. Linear relationship between mean damage ratings and Julian date in 1995, 1996

and 1997. The Lakes Country Club, Boiling Spring Lakes, Brunswick Co., NC. Julian dates presented as calendar dates for illustration.

149 LIST OF REFERENCES

Alexander, R. D. 1975. Natural selection and specialized chorusing behavior in insects. Pages 35–77 in Insects, science and society. Academic Press, New York. 284 pp.

Barbercheck, M. E. 1992. Effects of soil physical factors on biological control agents of soil insect pests. Fla. Entomol. 75: 539-548.

Barrett, O. W. 1902. The changa, or mole cricket (Scapteriscus didactylus Latr.) in Porto Rico. P. R. Agric. Stn. Bull. 2: 1-19.

Bennet-Clark, H. C. 1970. The mechanism and efficiency of sound production in mole crickets. J. Exp. Biol. 52: 619–652.

Brady, N. C. 1974. The nature and properties of soils. MacMillan, New York.

Braman, S. K. 1993. Progeny production, number of instars, and duration of development of tawny and southern mole crickets (Orthoptera: Gryllotalpidae). J. Entomol. Sci. 28(4):327–330.

Braman, S. K., and W. G. Hudson. 1993. Patterns of flight activity of pest mole crickets in Georgia. Intl. Turf. Soc. Res. J. 7:382–384.

Brandenburg, R. L., P. T. Hertl and M. G. Villani. 1997. Integration and adoption of a mole cricket management program in North Carolina, USA. Intl. Turf. Soc. Res. J. 8: 973- 979.

Brandenburg, R. L. and C. B. Williams. 1993. A complete guide to mole cricket management in North Carolina. NC Coop. Ext. Ser. ENT/ort - 101 8 pp.

Brandenburg, R. L., Y. Xia and A. S. Schoeman. 2002. Tunnel architecture of three species of mole crickets (Orthoptera: Gryllotalpidae). Fla. Entomol. 85(2): 383–385.

150 Cobb, P. P. and T. P. Mack. 1989. A rating system for evaluating tawny mole cricket, Scapteriscus vicinus Scudder, damage (Orthoptera: Gryllotalpidae). J. Entomol. Sci. 24(1):142–144.

Forrest, T. G. 1980. Phonotaxis in mole crickets: its reproductive significance. Fla. Entomol. 63(1): 45–53.

Forrest, T. G. 1981. Acoustic behavior, phonotaxis, and mate choice in two species of mole crickets (Gryllotalpidae: Scapteriscus). M S. thesis, Univ. of Florida, Gainesville. 72 p.

Forrest, T. G. 1983. Survival of eggs as a function of soil moisture. Ann. Rep. No.5, Mole Cricket Res. 82-83., Dept. of Ent. and Nem., University of Florida, Gainesville.

Forrest, T. G. 1983. Calling songs and mate choice in mole crickets. pp. 185-204. In D. T. Gwynne and G. K. Morris [eds.], Orthopteran mating systems: sexual competition in a diverse group of insects. Westview Press, Boulder, Colorado.

Forest, T. G. 1986. Oviposition and maternal investment in mole crickets (Orthoptera: Gryllotalpidae): effects of season, size and senescence. Ann. Entomol. Soc. Am. 79(6): 918–924.

Forrest, T. G. 1991. Power output and efficiency of sound production by crickets. Behav. Ecol. 2(4): 327-338.

Fowler, H. G. 1987. Geographic variation in flight periodicity of new world mole crickets. J. Interdiscipl. Cycle Res. 18(4): 283–286.

Fowler, H. G. 1989. Natural microbial control of populations of mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus borellii): regulation of populations aggregated in time and space. Rev. Brasil. Biol. 49(4): 1039–1051.

Hayslip, N. C. 1943. Notes on biological studies of mole crickets at Plant City, Florida. Fla. Entomol. 26(3): 33–46.

151 Henne, D. C. and S. J. Johnson. 2001. Seasonal distribution and parasitism of Scapteriscus spp. (Orthoptera: Gryllotalpidae) in southeastern Louisiana. Fla. Entomol. 84(2): 209–214.

Hertl, P. T., R. L. Brandenburg, and M. E. Barbercheck. 2001. The effect of soil moisture on ovipositional behavior in the southern mole cricket Scapteriscus borellii Giglio- Tos (Orthoptera: Gryllotalpidae). Environ. Entomol. 30: 466-473.

Hertl, P. T. and R. L. Brandenburg. 2002. Effect of soil moisture and time of year on mole cricket (Orthoptera: Gryllotalpidae) surface tunneling. Environ. Entomol. 31(3): 476–481.

Hertl, P. T., R. L. Brandenburg and C. B. Williams. 2003. Flight activity of Scapteriscus vicinus Scudder and S. borellii Giglio-Tos (Orthoptera: Gryllotalpidae) in southeastern North Carolina. (Manuscript in Preparation).

Hillel, D. 1980. Fundamentals of soil physics. Academic Press, New York.

Hinton, H. E. 1981. Biology of insect eggs. Pergamon Press, Oxford.

Hudson, W. G. 1985. Ecology of the tawny mole cricket, Scapteriscus vicinus (Orthoptera: Gryllotalpidae): Population estimation, spatial distribution, movement, and host relationships. Ph.D. dissertation, University of Florida.

Hudson, W. G. 1987. Variability in development of Scapteriscus acletus (Orthoptera: Gryllotalpidae). Fla. Entomol. 70(3): 403–404.

Hudson, W. G. 1989. Field sampling and population estimation of the tawny mole cricket (Orthoptera: Gryllotalpidae). Fla. Entomol. 72(2): 337–343.

Jessen, R. J. 1978. Statistical Survey Techniques. John Wiley and Sons. New York. 520 pp.

152 Kleyla, P. C. and G. Dodson. 1978. Calling behavior and spatial distribution of two species of mole crickets in the field. Ann. Entomol. Soc. Am. 71(4): 602–604.

Kramer, P. J. 1969. Plant and soil water relationships: a modern synthesis. McGraw-Hill, New York.

Krysan, J. L. 1976. Moisture relationships of the egg of the southern corn rootworm, Diabrotica undecimpunctata howardi (Coleoptera: Chrysomelidae). Entomol. Exp. Appl. 20: 154-162.

Lawrence, K. O. 1982. A linear pitfall trap for mole crickets and other soil arthropods. Fla. Entomol. 65(3): 376–377.

Marrone, P. G., and R. E. Stinner. 1983. Effects of soil moisture and texture on oviposition preference of the bean leaf beetle, Cerotoma trifurcata (Foster) (Coleoptera: Coccinellidae). Environ. Entomol. 12: 426-428.

Marrone, P. G., and R. E. Stinner. 1983. Effects of soil physical factors on egg survival of the bean leaf beetle, Cerotoma trifurcata (Foster) (Coleoptera: Chrysomelidae). Environ. Entomol. 12: 673-679.

Matheny, E. L., Jr. 1981. Contrasting feeding habits of pest mole cricket species. J. Econ. Entomol. 74(4): 444–445.

Matheny, E. L., Jr., R. L. Kepner and K. M. Portier. 1983. Landing distribution and density of two sound-attracted mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus). Ann. Entomol. Soc. Am. 76(2): 278–281.

Matheny, E. L. and B. Stackhouse. 1980. Seasonal occurrence and life cycles data for S. acletus and S. vicinus, field-collected in Gainesville, Florida. Ann. Rep. No. 2, Mole Cricket Research 79–80: 19–24.

Mihm, J. A., H. C. Chiang, and M. B. Windels. 1974. Moisture relationships of developing corn rootworm eggs, pp. 141-143. In Proceedings, Annual Meeting of the North Central Branch, Entomological Society of America, Lanham, MD.

153 Nevo, E., and A. Blondheim. 1972. Acoustic isolation in the speciation of mole crickets. Ann. Entomol. Soc. Am. 65: 980–981.

Ngo Dong, and H. W. Beck. 1982. Mark-release of sound-attracted mole crickets: flight behavior and implications for control. Fla. Ent. 65(4): 531–538.

Parkman, J. P., and J. H. Frank. 1992. Infection of sound-trapped mole crickets, Scapteriscus spp., by Steinernema scapterisci. Fla. Ent. 75(1): 163–165.

Parkman, J. P., and J. H. Frank. 1993. Use of a sound trap to inoculate Steinernema scapterisci (Rhabditida: Steinernematidae) into pest mole cricket populations (Orthoptera: Gryllotalpidae). Fla. Ent. 76(1): 75–82.

Parkman, J. P., W. G. Hudson, J. H. Frank, K. B. Nguyen, and G. C. Smart, Jr. 1993. Establishment and persistence of Steinernema scapterisci (Rhabditida: Steinernematidae) in field populations of Scapteriscus spp. mole cricket (Orthoptera: Gryllotalpidae). J Entomol. Sci. 28(2): 182–190.

Régnière, J., R. L. Rabb, and R. E. Stinner. 1981. Popillia japonica: Effects of soil moisture and texture on survival and development of eggs and first instar grubs. Environ. Entomol. 10: 654-660.

SAS Institute. 1992. SAS/STAT user’s guide, version 6.12. SAS Institute, Cary NC.

SAS Institute. 1999. SAS/STAT users guide, version 8.0. SAS Institute, Cary, NC.

SAS Institute. 2000. SAS/STAT user’s guide, version 8e. SAS Institute, Cary, NC.

Shetlar, D. J. and D. Herms. 1999. Insect and mite control on woody ornamentals and herbaceous perennials. Ohio State University Extension. Columbus OH. 73pp.

Short, D. E. and P. G. Koehler. 1977. Control of mole crickets in turf. Fla. Entomol. 60(2):147-148.

154 Short, D. E. and P. G. Koehler. 1979. A sampling technique for mole crickets and other pests in turfgrass and pasture. Fla. Entomol. 62(3):282–283.

Tsedeke, A. 1979. Plant material consumption and subterranean movements of mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus) as determined by radioisotopes techniques, with notes on materials for laboratory feeding. M.S. thesis. University of Florida, Gainesville. 72 pp.

Ulagaraj, S. M. 1975. Mole crickets: ecology, behavior, and dispersal flight (Orthoptera: Gryllotalpidae: Scapteriscus). Environ. Entomol. 4(2):265–273.

Ulagaraj, S. M. 1976. Sound production in mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus). Ann. Entomol. Soc. Am. 69(2): 299–306.

Ulagaraj, S. M., and T. J Walker. 1973. Phonotaxis of crickets in flight: attraction of male and female crickets to male calling songs. Science. 182: 1278–1279.

Ulagaraj, S. M., and T. J Walker. 1975. Responses of flying mole crickets to three parameters of synthetic songs broadcast outdoors. Nature. 253: 530–532.

Van Zwaluwenburg, R. G. 1918. The changa or West Indian mole cricket. Porto Rico Agric. Exp. Stn. Bull. 23: 1-28.

Villani, M. G. and R. J. Wright. 1990. Environmental influences on soil macroarthropod behavior in agricultural systems. Annu. Rev. Entomol. 35: 249-69.

Villani, M. G., L. L. Allee, L. Preston-Wilsey, N. Consolie, Y. Xia, R. L. Brandenburg. 2002. Use of radiography and tunnel castings for observing mole cricket (Orthoptera: Gryllotalpidae) behavior in soil. Am. Entomol. 48(1): 42–50.

Walker, S. L. 1979. Population estimation, damage evaluation and behavioral studies on the mole crickets Scapteriscus vicinus and S. acletus (Orthoptera: Gryllotalpidae). M.S. thesis. University of Florida, Gainesville. 83 pp.

155 Walker, T. J. 1982. Sound traps for sampling mole cricket flights (Orthoptera: Gryllotalpidae: Scapteriscus). Fla. Entomol. 65(1):105–110.

Walker, T. J. 1988. Acoustic traps for agriculturally important insects. Fla. Entomol. 7(4): 484–492.

Walker, T. J. and J. L. Nation. 1982. Sperm storage in mole crickets: fall mating fertilize spring eggs in Scpteriscus acletus. Fla. Entomol. 65(2): 283–285.

Walker, T. J., and G. N. Fritz. 1983. Migratory and local flight in mole crickets, Scapteriscus spp. (Gryllotalpidae). Environ. Entomol. 12(3): 953–958.

Walker, T. J. and D. A. Nickle. 1981. Introduction and spread of pest mole crickets: Scapteriscus vicinus and S. acletus reexamined. Ann. Entomol. Soc. Am. 74(2):158–163.

Walker, T. J., J. A. Reinert, and D. J. Schuster. 1983. Geographical variation in flights of the mole cricket, Scapteriscus spp. (Orthoptera: Gryllotalpidae). Ann. Entomol. Soc. Am. 76(3): 507–517.

Williams, J. J. and L. N. Shaw. 1982. A soil corer for sampling mole crickets. Fla. Entomol. 65(1): 192–194.

Worsham, E. L., and W. V. Reed. 1912. The mole cricket (Scapteriscus didactylus). Ga. Exp. Sta. Bull. 101: 251–263.

156