LISTENING TO THE LARVAE: ACOUSTIC DETECTION OF

304

Proc. Fla. State Hort. Soc.

116: 2003.

Jorgensen, J. (ed.). 1961. Fan Engineering. Buffalo Forge Company, Buffalo,

New York. 700 p.

Littell, R. C., P. R. Henry, and C. B. Ammerman. 1998. Statistical analysis of

repeated measures data using SAS procedures. J. Anim. Sci 76:1216-1231.

Littell, R. C., G. A. Miliken, W. W. Stroup, and R. D. Wolfinger. 1996. SAS Sys-

tems for Mixed Models, SAS Institute Inc., Cary, NC. 633 p.

Salyani, M. 1997. Performance of sprayers in Florida citrus production. 5th

Intl. Conf. on Fruit, Nut, and Vegetable Production Engineering, Davis,

Calif. 6 p.

Salyani, M. 2000. Methodologies for assessment of spray deposition in or-

chard applications. ASAE Paper No. 00-1031, 10 p., ASAE, St. Joseph,

Mich.

Salyani, M. and W. C. Hoffmann. 1996. Air and spray distribution from an air-

carrier sprayer. Appl. Eng. in Agr. 12:539-545.

Salyani, M. and J. D. Whitney. 1990. Ground speed effect on spray deposition

inside citrus trees. Trans. ASAE 33:361-366.

Salyani, M., L. W. Timmer, C. W. McCoy, and R. C. Littell. 2002. Petroleum

spray oils in Florida citrus applications, pp. 595-600. In A. Beattie, D. Wat-

son, M. Stevens, D. Rae, and R. Spooner-Hart (eds.). Spray Oils Beyond

2000 - Sustainable Pest and Disease Management.

SAS Institute, Inc. 1990. SAS User’s Guide: Statistics,

Ver.

6, SAS Institute,

Inc., Cary, N.C.

SPSS. 2000. SigmaPlot for Windows, Ver. 6, SPSS, Inc. Chicago, Ill.

Summerhill, W. R., J. L. Knapp, J. W. Noling, G. D. Israel, and C. L. Taylor.

1989. Citrus pest management. University of Florida, IFAS, Coop. Ext.

Ser. PE-6, 17 p.

Walklate, P. J., G. M. Richardson, and J. V. Cross. 1996. Measurements of the

effect of air volumetric flow rate and sprayer speed on drift and leaf de-

posit distribution from an air-assisted sprayer in an apple orchard. AgEng.

Madrid Paper 96A-131, 9 p.

Whitney, J. D. 1968. An analysis of application costs of airblast spraying in

Florida citrus. Proc. Fla. State Hort. Soc. 81:6-15.

Whitney, J. D., D. B. Churchill, S. L. Hedden, and R. P. Cromwell. 1986. Per-

formance characteristics of pto airblast sprayers for citrus. Proc. Fla. State

Hort. Soc. 99:59-65.

Whitney, J. D., M. Salyani, D. B. Churchill, J. L. Knapp, and J. O. Whiteside.

1989. A field investigation to examine the effects of sprayer type, ground

speed, and volume rate on spray deposition in Florida citrus. J. Agr. Eng.

Res. 42:275-283.

Proc. Fla. State Hort. Soc

. 116:304-308. 2003.

A REFEREED PAPER

LISTENING TO THE LARVAE:

ACOUSTIC DETECTION OF

DIAPREPES ABBREVIATUS

(L.)

ICHARD

W. M

ANKIN

USDA, ARS

1700 SW 23rd Drive

Gainesville, FL 32608

TEPHEN

L. L

APOINTE

USDA, ARS,

2001 South Rock Road

Ft. Pierce, FL 34945

Additional index words. Diaprepes abbreviatus

, citrus, root damage

Abstract.

Diaprepes abbreviatus

(L.) is an important pest of cit-

rus trees in Florida and the Caribbean. The larvae feed under-

ground on the root systems, reducing productivity and

facilitating invasion by root pathogens, including

Phytophtho-

spp. Field studies to survey or control larval populations

typically involve labor-intensive, destructive excavation of

root systems. However, nondestructive, portable instruments

are now available that can detect sounds made by insects

moving and feeding underground. Several different instru-

ments have been tested successfully for detection of subterra-

nean

D. abbreviatus

larvae and other insects, but many

questions remain about the use and reliability of acoustic de-

tection tools in specific insect-detection applications. This re-

port describes recent experiments with currently available

acoustic systems to assess the detectability and interpretabil-

ity of sounds produced by

D. abbreviatus

larvae and other or-

ganisms in root systems of individual trees in citrus groves. It

was confirmed that such instruments could successfully pre-

dict the presence or absence of subterranean insects under in-

dividual trees. However, the instruments do not provide a

comprehensive picture of the distribution of sounds (or in-

sects) around a tree unless multiple samples are recorded at

ca. 10-cm spacings within the root system. The rates of

sounds detected from a subterranean insect can vary consid-

erably at different times, depending on its patterns of behav-

ioral activity. The rates also can vary considerably at different

positions within a sensor’s detection range, depending on the

types of sound produced and the presence of roots or stones

between the insect and the sensor.

The development of improved methods for detecting and

monitoring

Diaprepes abbreviatus

(L.) (Coleoptera: Curculion-

idae) has been a long-standing priority for citrus growers and

regulatory agencies (

Diaprepes

Task Force, 1995; Nigg et al.,

1999). The larvae feed on the root systems of citrus trees (Bea-

vers and Selhime, 1975), directly damaging them or facilitat-

ing invasion by root pathogens, including

Phytophthora

spp.

(Rogers et al., 1996).

Until now, direct detection of the larvae usually has in-

volved labor-intensive excavation and inspection of root sys-

tems (McCoy et al., 2003). However, new acoustic

technologies are showing potential as rapid, nondestructive

tools for surveying subterranean insect infestations (Mankin

et al., 2000). Several different microphone and accelerome-

We thank Everett Foreman and Betty Weaver (Center for Medical, Agri-

cultural, and Veterinary Entomology, Gainesville, FL) for acoustic recording

and analysis and Anna Sara Hill and Laura Hunnicutt (U.S. Horticultural Re-

search Laboratory, Ft. Pierce, FL) for field assistance. Funds for this project

were made available from the Citrus Production Research Marketing Order

by the Division of Marketing and Development, Florida Department of Agri-

culture and Consumer Services, Bob Crawford, Commissioner.

Proc. Fla. State Hort. Soc.

116: 2003. 305

ter systems have been used successfully to detect sounds gen-

erated by subterranean larvae in citrus groves (Mankin et al.,

2001), forage fields (Brandhorst-Hubbard et al., 2001), and

nursery containers (Mankin and Fisher, 2002a). The success

of these initial studies led recently to the development of

user-friendly instrumentation designed specifically for insect

detection applications (Mankin and Fisher, 2002b).

Although considerable progress has been made in devel-

oping such instruments as insect detection tools, many unan-

swered questions remain about the interpretability of

measurements and the distances over which insects can be de-

tected. Such uncertainties are due partly to the physical struc-

ture of soil. Like stored grain (Shuman et al., 1997), soil is not

acoustically homogeneous. The presence of different soil tex-

tures, intervening roots, or stones can considerably alter the

temporal and spectral qualities of an insect-generated sound

over distances of a few cm. In addition, sounds produced at

the same location may be transmitted over considerably dif-

ferent distances if they are produced with different ampli-

tudes and spectral profiles (Mankin et al., 2000). As the

distance between a sound source and a sensor increases, un-

certainty about the original spectral profile and location of

the source increases considerably more rapidly in soil than in

air. We conducted a study in a citrus grove infested with

D. ab-

breviatus

in September, 2002, to gain further insights into the

detectability and interpretability of acoustic signals from sub-

terranean larvae in field environments.

Materials and Methods

Acoustic Systems.

Two portable acoustic systems were used

in this field study. The primary system, described in more de-

tail in Mankin et al. (2000, 2001) and Mankin and Fisher

(2002a), included an accelerometer and charge amplifier (0-

80 dB gain), a dual-channel digital audio tape recorder, and

a stereo headphone. The spectral range of the accelerometer

system was ca. 0-8 kHz. The accelerometer was attached to a

30-cm long, 0.6-cm-diameter steel probe, pushed into the soil

at an angle to pass near the crown of the citrus tree roots. The

amplified signal was monitored through the headphones and

passed to the recorder for further analysis (see

Signal Analysis

and Assessment of Infestation Likelihood

A second acoustic system, custom-designed for insect de-

tection applications (Mankin and Fisher, 2002b), included a

sensor-preamplifier module attached to a 20-cm-long, 0.6-cm-

diameter probe (Model SP-1, Acoustic Emission Consulting,

Inc. [AEC], Sacramento, Calif.). The preamplifier supplied

40-dB-gain between 1 and 50 kHz. The module was shielded

to reduce airborne background noise, and the reduced sensi-

tivity at frequencies <1 kHz eliminated much of the remain-

ing noise, which usually has peak frequencies below 400 Hz

(Mankin et al., 2000). The SP-1 sensor was attached to a Mod-

el AED-2000 (AEC, Sacramento, Calif.) amplifier unit provid-

ing a programmable, 0-60-dB additional gain. The amplifier

had an output for oscilloscopes or recorders, a headphone

port, a serial port for computer logging and signal display,

and a front-panel display of signal intensity and sound pulse

counts.

Recording and Infestation Verification Procedures.

Tests were

conducted with the accelerometer and SP-1 probes during

September 25-26, 2002, in an experimental grove of 15-yr-old

‘Minneola’ tangelo trees (

Citrus paradisi

Macf.

reticulata

Blanco) on

639 rootstock (

reticulata

Blanco

Poncirus tri-

foliata

(L.) Raf.)

at the IFAS Indian River Research and Edu-

cation Center, Ft. Pierce, Fla. The grove was on double beds

on Winder sand depressional soil (hyperthermic Typic Glos-

saqualfs). In an initial survey to locate sites with infestations

that could be studied in depth, a >3-min period was recorded

from one or more probes inserted into the soil underneath 10

separate trees. The probes were positioned within 10 cm of

the trunk and pointed towards the crown.

To consider the distribution of sounds around an entire

tree, recordings were done at 10 additional positions around

Tree No. 9, identified in the initial survey as one of the most

active trees (see

Signal Analysis and Assessment of Infestation

Likelihood

). Several of these measurements were made by si-

multaneously feeding the output from the accelerometer am-

plifier and the AED-2000 into separate channels of the dual

recorder, enabling the same sounds to be compared instanta-

neously at multiple positions.

Just before excavation on the second day, the tops of

Trees No. 6,7, 9, and 10 were cut off at ca 10-cm height and

simultaneous recordings were obtained from probes inserted

into the trunk and in the soil. The tops were removed to re-

duce interference from wind noise (Mankin et al., 2000,

2002) and facilitate direct comparison of the acoustic signals

from the trunk and the soil probes. Loud sounds were detect-

ed both in the soil and in the trunk at Tree No. 6. For an in

depth analysis, we first recorded three consecutive 3-min in-

tervals with the SP-1 probe inserted into the trunk and the ac-

celerometer probe in the soil at a single position, 3 cm from

the trunk. The signals were fed simultaneously into separate

channels of the recorder. A fourth interval then was recorded

with the SP-1 probe moved to a point 3 cm from the trunk and

3 cm from the accelerometer probe.

After the recordings were completed on the second day,

the tested trees were excavated and the root systems were ex-

amined. The numbers of

D. abbreviatus

larvae and other

sound-producing organisms recovered from the root systems

were noted for comparison with the acoustic assessments of

infestation likelihood (see next section).

Signal Analysis and Assessment of Infestation Likelihood.

The

recorded signals were digitized in the laboratory and quanti-

tatively analyzed using a custom-written signal processing sys-

tem (Mankin et al., 2000, 2001). When high levels of

background noise sometimes interfered with analysis of a

complete 3-min recording, we evaluated shorter, contiguous

sections; however, the file was discarded if we could not find

an analyzable section of at least 30-s.

Moving and feeding

D. abbreviatus

grubs produce 2-5-ms

clicks and scraping sounds that experienced listeners and

computer programs identify as grub sound pulses (Mankin et

al., 2001; see also http://cmave.usda.ufl.edu/~rmankin/

soundlibrary.html). These distinctive pulses are easily distin-

guishable from wind noise, bird calls, or engine noise, but

cannot be distinguished easily from sounds made by mole

crickets (

Scapteriscus

sp.) and other subterranean insects often

found in citrus groves (Mankin et al., 2000). In the rest of this

report, all such sounds were designated as grub pulses, al-

though some of them were produced by

Scapteriscus

(see Ta-

ble 1). The rate of grub sound pulses can be used as a guide

to assess the likelihood that a tree is infested. In a previous

study of

D. abbreviatus

(Mankin et al., 2001), the likelihood of

infestation was scaled as

low

for pulse rates

≤

2/min,

medium

for rates between 2 and 20/min, and

high

for rates > 20/min.

We adopted the same likelihood scale in this study.

306

Proc. Fla. State Hort. Soc.

116: 2003.

Results and Discussion

Accelerometer probes under Trees No. 1, 6, and 9 detect-

ed signals with 2-5 ms durations and spectral peaks between

600 and 2000 Hz that are typical of

D. abbreviatus

grub sound

pulses (Mankin et al., 2001). The likelihood of pest infesta-

tion was rated

medium

at Tree No. 1, where sounds were de-

tected at a low rate of 7/min (Table 1). The likelihood was

rated

high

at Tree No. 6 and No. 9, where rates were >20/min

in many recordings (columns 2-3 in Table 2 and column 5 in

Table 3). The distribution of actual infestations is compared

with the distribution of predicted infestations in Table 4. In

this case, the mole cricket at Tree No. 6 was counted as a pest

because mole crickets are considered harmful to root systems.

The relationship between computer-rated likelihood and the

observed infestation was statistically significant (

= 6.43, 2

df,

< 0.05). The results were similar to those in Mankin et al.

(2001), where highly active insects were quickly detected but

the absence of sound could indicate either that no insect was

present or that no insect was active.

Comparisons among SP-1 and Accelerometer Sensors.

From pre-

vious experience, we expected that different sounds pro-

duced by

D. abbreviatus

larvae might be detected differentially

by the SP-1 and accelerometer systems, depending on the

spectral patterns of the sounds and the distance between the

probe and larva. Grub sound pulses have highly variable am-

plitudes and spectral patterns, partly because different feed-

ing and movement activities generate sounds with different

spectral profiles (Mankin et al., 2000, 2001). Root clipping

and scraping sounds, for example, might be detected most

easily by a SP-1 probe inserted into the trunk. These sounds

have high frequency components that travel well in wood (see

Mankin et al. 2002 and references therein), and SP-1 probes

are differentially sensitive to high frequencies. Low-frequen-

cy, low-amplitude sliding movements might be detected most

easily by the accelerometer because it is highly sensitive to sig-

nals <1 kHz that travel well in soil (Mankin et al., 2000). How-

ever, many insect sounds contain both high- and low-

frequency components that might be detected by both sen-

sors, especially if the insect were nearby and the signals were

of high amplitude.

We confirmed in a series of recordings at Tree No. 6 that

both sensors could detect the same nearby, high-amplitude

sound source. On day 2, we detected frequent pulses at the

first position tested with the accelerometer probe, ca. 3 cm

from the trunk. The high-amplitude of the signal indicated

that an insect was nearby, next to the trunk. We took three

consecutive, 3-min simultaneous recordings with the acceler-

ometer at its original position and the SP-1 probe in the trunk.

A fourth simultaneous measurement was taken immediately

afterward, with the SP-1 probe in the soil, 3 cm from the accel-

erometer and the trunk (Table 2). When the tree was extract-

Table 1. Numbers of subterranean organisms recovered and rates of grub

sound pulses detected by accelerometer under tangelo trees in

Diaprepes root weevil-infested grove.

Tree

No.

D. abbreviatus

No.

other insects

No. grub-sound

pulses/min

131 (

Scapteriscus

sp.) 7

23 0

01 (

Scapteriscus

sp.) see Table 2

31 (

Dermaptera

sp.) see Table 3

After subterranean recordings were completed, the top of this tree was

removed and recordings were made from a probe inserted into the top of

the stump.

Table 2. Rates of grub sound pulses in three consecutive 3-min intervals with

the accelerometer (ACC) at a single position, 3 cm from the trunk of

Tree No. 6, recorded simultaneously with the SP-1 probe inserted into

the trunk, followed by one recording with SP-1 in soil, 3 cm from ACC.

No. grub sound pulses/min dB

Test No. ACC SP-1 ACC SP-1

6.1 127.3 21.1 18.0 10.2

6.2 212.2 86.7 20.3 16.4

6.3 236.3 232.0 20.7 20.6

6.4 243.7 115.0 20.8 17.6

No. grub sound pulses/min transformed into decibel scale using: dB = 10

log

(pulse rate/

), where

= 2 pulses/min (see

Comparisons among SP-1

and Accelerometer Sensors

Table 3. Detection rates of grub sound pulses recorded with accelerometer (ACC) or SP-1 probes at multiple positions under Tree No. 9.

Test No. Position

Probe Recorded duration (sec) No. grub sound pulses/min dB

9.1 P1 ACC 122 8.9 6.5

9.2 P2 ACC 180 6.3 5.0

9.3 P3 ACC 180 12.3 7.9

9.4

ACC 86 5.6 4.5

9.5, 9.6 P4, P5 ACC 180 0.0 —

9.7 P6 ACC 180 6.0 4.8

9.8 P7 SP-1 180 48.0 13.8

9.9

SP-1 172 89.7 16.5

9.10 P8 ACC 60 48.0 13.8

9.11

ACC 60 9.0 6.5

9.12 P9 ACC 180 0.0 —

9.13 Trunk SP-1 180 108.0 17.3

See Fig. 1 for details of spacing around tree.

See Table 2.

Second recording obtained ca. 3 min after the previous recording at same site.

Proc. Fla. State Hort. Soc.

116: 2003. 307

ed and the roots examined, the sound source was found to be

a single mole cricket, (see sample posted at http://cmave.

usda.ufl.edu/~rmankin/molecricket4-sept02.wav). Mole

cricket sounds cannot yet be reliably distinguished from

D. abbreviatus (

Fig. 1).

As expected, the measurements with the two probes pro-

vided essentially equivalent results in the paired recordings.

The mean rate of sounds detected by the SP-1 probe, 113.7 ±

44.1 grub pulses per min, was less than the rate detected by

the accelerometer, 204.9 ± 26.7 grub pulses per min, but the

difference was not statistically significant (

= 1.76, df = 4,

0.14). Because the sample variability was large and standard

errors were not homogeneous, it was convenient to transform

the results using a decibel scale based on the threshold distin-

guishing

low

from

medium

infestation likelihood,

= 2 grub

sounds/min (Mankin et al., 2001). The dB-transformed val-

ues are listed in the last two columns of Table 2. Considering

that the difference between 244 and 115 pulses per min is

probably less relevant behaviorally than the difference be-

tween 127 and 21 pulses per min, the transformed values are

easier to interpret.

Comparisons among Closely Spaced Sensor Positions.

In multi-

ple recordings under Tree No. 9, grub sound pulses were de-

tected at eight of ten positions (Fig. 1), and the rate of sound

production varied from background levels up to 234 grub

pulses per min (Test No. 9.16 in Table 5). The high rates of

sounds detected at positions,

and

, suggests that at least

one of the four recovered insects (see Table 1) was located be-

tween them. As at Tree No. 6, the rates varied considerably in

consecutive recordings at the same position (Tests No. 9.8-

9.9, and 9.16-9.17 at

, Tests No. 9.10-9.11 at

, and Tests

No. 9.12 and 9.17 at

). Consequently, the decibel values list-

ed in the last two columns may be more relevant for temporal

comparisons than the untransformed rates. High variability

also was observed in the rates of sounds detected from closely

spaced sensors. In Test No. 9.15 of Table 5, for example,

sounds detected at high rates were barely detectable in a si-

multaneous recording at a probe 11 cm away. Such variability

can be of utility in locating individual insects, but to map out

a comprehensive distribution of the locations and sound pat-

terns of all the insects around a tree would be more time- and

labor-intensive than is usually feasible.

The high spatial and temporal variability of the sound rates

that we have observed in this and related studies currently lim-

its their utility for quantitative analysis of subterranean insect

behavior. It is not realistic to expect that many field studies will

be performed with multiple recordings from probes spaced 10-

cm apart throughout the root systems of multiple trees. In the

future, however, we hope to develop improved sound classifi-

cation methods that may help distinguish among different lar-

val behaviors, and improve the capability to distinguish among

different species. In that case, measurement of the rates of spe-

cific types of sound may provide additional information that

cannot be easily be obtained solely from the unadjusted total

rate of grub sound pulses. Even without such improvements,

the recent development of more portable, user-friendly instru-

mentation has increased the opportunity of researchers and

grove managers to use acoustic technology in a variety of new

D. abbreviatus

detection applications.

Table 4.

Numbers of uninfested and (

D. abbreviatus

Scapteriscus

) pest-

infested trees rated at different likelihoods of pest infestation by com-

puter analysis of grub sound pulse rates.

Computer-rated

infestation likelihood

No.

uninfested trees

No.

infested trees

Low 6 1

Medium 0 2

High 0 2

Basis of computer-rated infestation likelihood:

low

≤

2 grub pulses/min;

medium

, 20

≤

rate > 2 grub pulses/min;

high

, >20 grub pulses/min.

Fig. 1. Mean rates of detection of grub sound pulses recorded from differ-

ent positions with accelerometer or SP-1 probe in soil underneath citrus tree

or inserted into trunk (see Table 3).

Table 5. Detection rates of grub sound pulses in simultaneous recordings with accelerometer (ACC) and SP-1 probes at multiple positions under Tree No. 9.

Sensor Probe Position No. pulses/min dB

Test No. Probe 1 Probe 2 Probe 1 Probe 2 Dist

. (cm) Probe 1 Probe 2 Probe 1 Probe 2

9.14 ACC ACC P3 P6 6 5.0 285.7 4.0 21.5

9.15 SP-1 ACC P7 P6 4 40.7 8.0 16.1 6.0

9.16 SP-1 ACC P7 P10 11 234.0 6.0 20.7 4.8

9.17 ACC ACC P9 P10 4 8.0 1.3 4.8 —

Distance between probes in cm.

See Table 2.

308

Proc. Fla. State Hort. Soc.

116: 2003.

Literature Cited

Beavers, J. B. and A. G. Selhime. 1975.

Biology of

Diaprepes abbreviatus

(Co-

leoptera: Curculionidae) reared on an artificial diet. Fla. Entomol.

65:263-269.

Brandhorst-Hubbard, J. L., K. L. Flanders, R. W. Mankin, E. A. Guertal, and

R. L. Crocker. 2001. Mapping of soil insect infestations sampled by exca-

vation and acoustic methods. J. Econ. Entomol. 94:1452-1458.

Diaprepes

Task Force. 1995. Florida Department of Agriculture and Consum-

er Services, Division of Plant Industry, Bureau of Pest Eradication and

Control, Gainesville, FL.

Mankin R. W. and J. R. Fisher. 2002a.

Acoustic detection of black vine weevil,

Otiorhynchus sulcatus

(Fabricius) (Coleoptera: Curculionidae) larval infes-

tations in nursery containers. J. Environ. Hort. 20:166-170.

Mankin, R. W. and J. R. Fisher. 2002b. Current and potential uses of acoustic

systems for detection of soil insect infestations, pp. 152-158. In: Proc

Bouyocous Conference on Agroacoustics, Fourth Symposium, National

Center for Physical Acoustics, University, MS, May 6-9, 2002.

Mankin, R. W., J. Brandhorst-Hubbard, K. L. Flanders, M. Zhang, R. L.

Crocker, S. L. Lapointe, C. W. McCoy, J. R. Fisher, and D. K. Weaver.

2000. Eavesdropping on insects hidden in soil and interior structures of

plants. J. Econ. Entomol. 93:1173-1182.

Mankin, R. W., S. L. Lapointe, and R. A. Franqui. 2001. Acoustic surveying of

subterranean insect populations in citrus groves. J. Econ. Entomol.

94:853-859.

Mankin, R. W., W. L. Osbrink, F. M. Oi, and J. B. Anderson. 2002. Acoustic

detection of termite infestations in urban trees. J. Econ. Entomol. 95:981-

988.

McCoy, C. W., R. J. Stuart, and H. N. Nigg. 2003. Seasonal life stage abun-

dance of

Diaprepes abbreviatus

in irrigated and non-irrigated citrus plant-

ings in central Florida. Fla. Entomol. 86:34-42.

Nigg, H. N., S. E. Simpson, L. E. Ramos, and N. W. Cuyler. 1999. Assessment

of monitoring techniques for

Diaprepes abbreviatus

(L) Coleoptera: Curcu-

lionidae). Proc. Fla. State Hort. Soc. 112:73-77.

Rogers, J. S., C. W. McCoy, and J. H. Graham. 1996.

Insect-Plant pathogen in-

teractions: preliminary studies of

Diaprepes

root weevil injuries and

Phy-

tophthora

infection. Proc Fla. State Hort. Soc. 109:57-62.

Shuman, D., D. K. Weaver, and R. W. Mankin. 1997. Quantifying larval infes-

tation with an acoustical sensor array and cluster analysis of cross-correla-

tion outputs. Appl. Acoust. 50:279-296.

Proc. Fla. State Hort. Soc

. 116:308-311. 2003.

A REFEREED PAPER

TRIFLOXYSULFURON-SODIUM – A POSSIBLE NEW HERBICIDE

FOR WEED CONTROL IN CITRUS

HIV

D. S

HARMA

AND

M. S

INGH

University of Florida, IFAS

Citrus Research and Education Center

700 Experiment Station Road

Lake Alfred, FL 33850-2299

Additional index words. adjuvant, CGA-362622, Envoke®, gly-

phosate, guineagrass, halosulfuron, nutsedge, organosilicone,

pigweed, Spanish needle, sulfonylurea, surfactant

Abstract. Bioefficacy studies of trifloxysulfuron-sodium, a new

sulfonylurea herbicide, were conducted with and without sur-

factant and compared with glyphosate (Rodeo) and halosulfu-

ron-methyl (Permit). In general, trifloxysulfuron-sodium was

not effective without a surfactant, except the highest rate (31.5

g a.i./ha). Application of trifloxysulfuron-sodium with adju-

vants (nonionic – X-77, organosilicone – L-77 or oil – MSO) sig-

nificantly increased herbicide efficacy. Application of

glyphosate at 500 g a.i./ha

increased weed mortality signifi-

cantly by providing 51%, 65%, 83%, and 88% control of yellow

nutsedge (Cyperus esculentus L.), guineagrass (Panicum

maximum Jacq.), redroot pigweed (Amaranthus retroflexus

L.), and hairy Spanish needles (Bidens bipinnata L.), respec-

tively. With the exception of guineagrass, increasing dosage

of halosulfuron (8.75 to 35 g a.i./ha) did not influence yellow

nutsedge, redroot pigweed, or Spanish needles. Trifloxysulfu-

ron-sodium, even at 7.5 g a.i./ha, was comparable with the

highest rates of glyphosate (500 g a.i./ha) or halosulfuron (35

g a.i./ha). Trifloxysulfuron-sodium at 30 g a.i./ha provided max-

imum control (≥86%) of all the test weed species. In subse-

quent studies, application of trifloxysulfuron-sodium to three

different citrus rootstocks resulted in significant phytotoxic

effects to the primary stem in the form of necrotic leaves and

further growth was stopped. However, upon pruning the ne-

crotic tissue, lateral growth arose from the trimmed point fol-

lowed by normal rootstock growth.

The environment is subjected to a greater risk when high

rates of pesticides are used for pest control. In addition, re-

petitive use of a single active ingredient over time increases

chances of resistance development in the target pests. Gly-

phosate, which has been used worldwide for more than 20

years (Bradshaw et al., 1997) and accounts for 11% of the

worldwide sales of herbicide (Powles et al., 1997) is a good ex-

ample. Continuous use of glyphosate has resulted in the ap-

pearance of resistant weed populations of rigid ryegrass

(Lolium rigidum Goud.) (Holt et al., 1993; Powles and Hol-

tum, 1994; Powles et al., 1998).

Trifloxysulfuron-sodium [N-(4,6-Dimethoxy-2-pyrimidi-

nyl)-3-(2,2,2-trifluoroethoxy)-pyridin-2-sulfonamide sodium

salt] is a new broad-spectrum, low-rate technology herbicide

for over-the-top post-emergence application, developed for

use in sugarcane and cotton (Rawls et al., 2000). This sulfony-

lurea herbicide, with the proposed common name of triflox-

ysulfuron sodium, has been field tested as a 75% water-

dispersible granule (WDG) for the past several years in North

America, Africa, and Asia under the code name trifloxysulfu-

The authors thank Syngenta Crop Protection for research samples of her-

bicides and financial help in conducting the experiment and Gary Test for

help and maintenance of the experiments. This research was supported by

the Florida Agricultural Experiment Station, and approved for publication as

Journal Series No. R-08821.

Corresponding author; e-mail: [email protected]. edu.