Citation
Sampling programs for and ecological relationships of Frankliniella spp. (Thysanoptera:thripidae) on tomato

Material Information

Title:
Sampling programs for and ecological relationships of Frankliniella spp. (Thysanoptera:thripidae) on tomato
Creator:
Salguero Navas, Victor Eberto, 1952-
Publication Date:
Language:
English
Physical Description:
x, 114 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Cotton ( jstor )
Crops ( jstor )
Density estimation ( jstor )
Flowers ( jstor )
Infestation ( jstor )
Insects ( jstor )
Pests ( jstor )
Species ( jstor )
Tomatoes ( jstor )
Wildlife management ( jstor )
Dissertations, Academic -- Entomology and Nematology -- UF ( lcsh )
Entomology and Nematology thesis Ph. D ( lcsh )
Gadsden County ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 102-113).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Victor Eberto Salguero Navas.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
023318823 ( ALEPH )
23011913 ( OCLC )

Downloads

This item has the following downloads:


Full Text












SAMPLING PROGRAMS FOR AND ECOLOGICAL RELATIONSHIPS
OF Frankliniella SPP. (THYSANOPTERA:THRIPIDAE) ON TOMATO.














By

VICTOR EBERTO SALGUERO NAVAS


















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1990












ACKNOWLEDGEMENTS

I wish to express my sincere gratitude to Dr. Joseph E. Funderburk, Dr. Donald C. Herzog, Dr. Richard K. Sprenkel, and Dr. Steve Olson for serving on the advisory committee and for giving of their time and experience to improve my professional development in integrated pest management. Their constructive criticisms were invaluable in the course of this research and in the preparation of this manuscript.

My gratitude is especially extended to Dr. Ramona Beshear, University of Georgia, for generously sharing her time and skills on the identification of thrips and to Dr. Timothy P. Mack, Auburn University, for his unconditional assistance in the statistical analysis of data.

Sincere thanks are also due to the personnel of the North Florida Research and Education Center in Quincy and the Department of Entomology and Nematology in Gainesville. Especial thanks go to Myrna Litchfield, Sheila Eldrige, Tracey Austin, Jan Smith, Jan Gray, Janice Walden, Connie Rudd, Elizabeth Lewis, Andrew Brown, Glen Porcieau, and others for their endless patience and willingness to assist me during the development of this study.


ii








Thanks are extended to Dr. Avas Hamon and Dr. Charles Niblett for giving me advice and moral support whenever needed. Their warm friendship and exemplary manners will be always in my life.

I wish to acknowledge to the Instituto de Ciencia Y Tecnologia Agricolas in Guatemala and its authorities for providing me with the scholarship to get my Ph.D. and especially for allowing me to properly orient my profesional development.

To my loving daughter Cindy, my friend and son Juan Carlos, and my beautiful wife Lissette, I give thanks for their patience and confidence in my abilities. Their presence has made my hardships seem insignificant.

Thanks my parents, Jose Victor and Maria Luisa, whose example of love and constant support were the basis for all I have achieved. They gave me more than I can ever repay.



















iii











TABLE OF CONTENTS



ACKNOWLEDGEMENTS ................*.*.* ..*.*.*........ ii

LIST OF TABLES ... ............. **********....******** vi

LIST OF FIGURES .......................**.*.**.......*.... vii

ABSTRACT .................. ******.** * .* * .* * ....... ix

CHAPTER
1. INTRODUCTION ................................. 1

2. LITERATURE REVIEW ............................. 6

Insects and related pests in tomato ........... 6
Flower thrips Frankliniella spp .............. 8
General structure and systematics of the
order Thysanoptera ..................... 8
Biology of Frankliniella spp. .............. 15
Ecology of Frankliniella spp. .............. 17
Host-plant relationships ............... 17
Movement, migration, and dispersal ..... 19 Survival and natural regulation ........ 20
Economic importance ........................ 21
As pests on several crops ............. 21
As predators of other pest species ..... 24 Pollination of flower .................. 24
Tomato spotted wilt virus (TSWV) .............. 25
Host range of TSWV ........................ 26
TSWV transmission .......................... 26
Symptoms of TSWV ......................... 28
Management of Frankliniella spp. and TSWV ..... 30
Pest management in tomatoes ................ 30
Control of Frankliniella spp. and TSWV ..... 33 Surveillance and sampling programs ......... 38
Thrips sampling techniques ............. 39
Spatial distribution patterns .......... 43 Number of samples ...................... 46
Seasonal distribution .................. 47
Binomial (or presence/absence)
sampling program .................. 49
Factors affecting estimation of
insect densities .................. 51
Economic injury level and economic thresholds . 53


iv












3. SEASONAL PATTERNS OF Frankliniella SPP.
(THYSANOPTERA:THRIPIDAE) IN TOMATO FLOWERS
AND INFLUENCES OF SEVERAL FACTORS ON SAMPLE
ESTIMATES ........... ....... ........... ....... 57

Introduction .................................. 57
Materials and Methods ......................... 58
Results and Discussion ........................ 61

4. BINOMIAL SAMPLING PROGRAM FOR Frankliniella
SPP. IN TOMATO ACCORDING TO THE POSITION OF
SAMPLING IN THE PLANT ......................... 73

Introduction ................................. . 73
Materials ans Methods ......................... 75
Results and Discussion ....................... 77

5. FLOWER THRIPS, Frankliniella SPP. DAMAGE TO
TOMATO THRIPS ......................... ........ 85

Introduction .................................. 85
Materials and Methods ......................... 86
Results and Discussion ........................ 88

6. CONCLUSIONS ................ ................... . 94

REFERENCES ........................................... 102

BIBLIOGRAPHICAL SKETCH ............................... 114



















v

















LIST OF TABLES


Table 2.1. Key to the adult females of Frankliniella
spp. thrips found in tomato flowers ........ 12

Table 2.2. Developmental rate of Frankliniella spp.
thrips ..................................... 18

Table 3.1. Effects of field position, plant position,
and time of sampling on seasonal densities of adult E. occidentalis, E. fusca, and E.
triticiand total adult and immature thrips
in flowers in two tomato fields sampled during the spring 1988 and 1989 growing
seasons in Gadsden Co., Florida ............ 72

Table 4.1. Regression statistics of Taylor's power
law relationships for Frankliniella spp.
thrips sample data in tomato fields in
North Florida during 1988 and 1989 ......... 82

Table 5.1. The mean number of scars per mature green
tomato fruit resulting from female
Frankliniella spp. thrips confined on
flowers and small fruit .................... 92











vi










LIST OF FIGURES

Fig. 2.1. Key to the adult females of Frankliniella
thrips found in tomato flowers ............. 14
Fig. 3.1. Mean number of adult E. occidentalis, E.
tritici and �. fusca per flower in tomato
fields sampled weekly during the spring
of 1987, 1988, and 1989 in Gadsden County,
Florida .................................... 66

Fig. 3.2. Mean number of adult E. occidentalis and
E. tritici per flower in tomato fields sampled weekly during the fall of 1987
and 1988 in Gadsden County, Florida ........ 67

Fig. 3.3. Mean number of flowers and adult thrips
per plant in the tomato fields sampled
weekly during the springs of 1988 and
1989 and the fall of 1988 in Gadsden
County, Florida ............................ 68

Fig. 3.4. Effect of field position of sampling on
density estimates of Frankliniella spp.
thrips in flowers in two tomato fields
sampled during the spring in Gadsden
County, Florida ............................ 69

Fig. 3.5. Effect of plant position of sampling on
density estimates of Frankliniella spp.
thrips in flowers in two tomato fields
sampled during the spring in Gadsden
County, Florida ............................ 70

Fig. 3.6. Effect of time of sampling on density
estimates of Frankliniella spp. thrips
in flowers in two tomato fields sampled
during the spring in Gadsden County,
Florida .................................... 71

Fig. 4.1 The relationships between the proportion
of infested flowers and Frankliniella
spp. density per tomato flower as
estimated by using the Wilson and Room
(1983) equation ........................... .. . 83

vii








Fig. 4.2. The relationships between the number of
samples needed to estimate density at
the 10 and 25% precision levels and the
number of thrips per flower for
Frankliniella spp. in tomato as determined
by using the Ruesink (1980) equation ........ 84

Fig. 5.1. Relationship between female E. occidentalis
density and the number of scars per tomato
fruit ....................................... 93





































viii










Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy


SAMPLING PROGRAMS FOR AND ECOLOGICAL RELATIONSHIPS
OF Frankliniella spp. (THYSANOPTERA:THRIPIDAE) IN TOMATO

By
Victor Eberto Salguero Navas

May 1990

Chairman: J. E. Funderburk
Co-chairman: D. C. Herzog
Major Department: Entomology and Nematology


Flower thrips (Frankliniella spp.) densities in tomato flowers were estimated weekly in commercial fields during the growing seasons of 1987, 1988, and 1989. Thrips were abundant during the spring but almost absent in the fall. E. occidentalis (Pergande), E. tritici (Fitch), and E. fusca(Hinds) were commonly collected during the spring. Densities were greatest during May of each year. Species collected in the fall were E. tritici and �. occidentalis.

The effects of several factors on thrips density estimates were investigated. Plant position of sampling had a significant effect on density estimates of E. occidentalis and F. tritici, but not E. fusca. Adult thrips of all species were more abundant in flowers located on the upper half of the plants than on flowers located on the lower half of the plants. Immatures were more abundant on the lower ix








half. Sample location within a field did not affect density estimates of most thrips species. However,density estimates of E. occidentalis in 1988 were greater in tomato flowers located near margins than in flower located in nonmarginal areas. A similar tendency was observed in 1989 but this effect was not statistically significant. The time of day of sampling had no effect on thrips density estimates.

Dispersion characteristics were quantified by using the Taylor's power law relationships for adult species and immatures. Most adult species and immatures were aggregated over a wide range of densities. E* fusca populations were slightly aggregated to random. A binomial sampling program was developed for adult E. occidentalis, E. tritici, and �. fusca inhabiting the upper half of tomato plants. The slope and intercept of the Taylor's power law relationships were used to determine the relationships between proportion of infested flowers and density. The number of samples needed to estimate thrips densities at the 10 and 25% precision levels was also determined.

Cosmetic fruit damage was found to be caused by oviposition from E. occidentalis. The other species were not found to cause cosmetic damage. The relationships between the number of female thrips in tomato flowers and the number of scars on the fruit was very variable. This is expected to hamper efforts to develop economic injury levels for E. occidentalis in tomatoes.

x













CHAPTER 1

INTRODUCTION

Flower thrips guild, Frankliniella spp., constitutes one of the most complex polyphagous pest groups affecting the production of many ornamental and crop plant species. They are very general feeders (virtually cosmopolitan) on many plant species belonging to diverse families (Eddy and Livingstone 1931, Watts 1936, Newsom et al. 1953, Bryan and Smith 1956, Graves et al. 1987). The western flower thrips, E. occidentalis (Pergande), is considered an important pest in cotton (Stoltz and Stern 1978a, b, Rummel and Quisenberry 1979, Wilson 1982, Graves et al. 1987, Pickett et al 1988), table grapes (Yokoyama 1977), safflower (Carlson 1964b, 1966, Carlson and Witt 1977), onions (Harding 1961, Carlson 1964a, Dintenfass et al. 1987), many ornamentals (Baker and Stephan 1986, Jones and Moyer 1986), apples (Terry and DeGrandiHoffman 1988), and tomato (Oetting 1985, Olson and Funderburk 1986, Cho et al. 1989).

The tobacco thrips, E. fusca (Hinds), has been reported attacking peanuts (Morgan et al. 1970, Tappan and Gorbet 1979, 1981, Lynch et al. 1984, Tappan 1986a, b), and cotton (Watts 1934, 1937a, b, Watson 1965, Rummel and Quisenberry 1979). The flower thrips, E. tritici (Fitch), has been reported from



1








2

commercial roses (Stannard 1968), soybean (Marston et al. 1979), cotton (Watts 1934, Watson 1965), and peanuts (Morgan et al. 1970). E. bispinosa (Morgan) (there is no accepted common name) has been reported from peanuts (Morgan et al. 1970).

Flower thrips are important pests in tomato due to direct damage to the fruit which results in superficial or cosmetic damage which is not acceptable in a fresh market product by consumers or by national federal standard. Additionally, E. occidentalis and E. fusca are vectors of tomato spotted wilt virus (TSWV) which seriously affects production of tomato and other food and ornamental crops worldwide (Cho et al. 1989). In North America, TSWV has caused economic losses in Canada (Allen and Broadbent 1986), Hawaii (Cho et al. 1986), and Louisiana (Greenough et al. 1985).

Tomato is economically the most important vegetable crop grown in Florida (Pohronezny et al. 1986). Almost 98% of Florida tomatoes are grown for fresh market sales. Of the $711 million annual total U.S. farm value for fresh market tomatoes, Florida accounted for 53% ($377 million) (Clough 1987). The total value of this crop in Florida increased from $377 millions in 1987 to $603 millions in the season of 19881989 (Anonymous 1989).

This crop is characterized by having a few major key pests and many secondary or minor pests whose relative








3

importance varies both within and among major production areas. These pests can cause an economic reduction in optimum yield or damage the fruit which in fresh market tomatoes reduces the effective marketable yield because even slightly

damaged fruit are rendered unmarketable. Deformation of fruit may be produced by the feeding of caterpillars, aphids, or thrips (Lange and Bronson 1981).

Neither the flower thrips complex nor TSWV have been considered as important tomato pests in Florida in the past (Pohronezny et al. 1986). E. tritici, E. fusca, E.bispinosa, and E. cephalica masori Wats were reported in Florida as early as 1913 (Watson 1922, 1923). E. occidentalis is not considered to be native to the Southeast; its natural range in the U.S. is the West-North regions and Canada (Baker and Stephan 1986). However, Watson (1923) reported this species in South Florida as early as 1910. The first published report of this species from the Southeast was from Georgia on cotton flowers in 1981 (Beshear 1983). However, it had been collected on soybean in North Carolina in 1978 (Baker and Stephan 1986). It was later found for the first time in cotton in Louisiana in 1985 (Graves et al. 1987).

Direct damage of Frankliniella spp. to tomato fruit was first recognized in North Florida as an economic problem in 1985 about the time E. occidentalis was first recorded in that area (Funderburk 1988). In addition to E. occidentalis, E. fusca, E. tritici and E. bisDinosa have been found to inhabit








4

tomato blooms (Salguero and Funderburk 1989). The cosmetic injury produced by flower thrips on tomato fruit has become so commonplace that rejections of fruit by regulatory authorities has already occurred (personal observations).

TSWV, on the other hand, was first detected and confirmed in tomato in North Florida in 1986 (Sprenkel 1988). It was also found on peanuts in the same area in 1989 (Sprenkel 1989). The incidence of this disease in tomato fields increased from less than 0.5% in 1988 (Sprenkel 1988) to 2% in 1989 (Helena Puche, unpublished data). The presence of TSWV and Frankiniella spp. in North Florida is causing alarm among tomato producers causing them to increase both the dosages and frequency of insecticide applications. Insecticidal control measures usually suppress the adult thrips populations; however, they rapidly rebound following treatment because of migration into the field. This migratory characteristic, polyphagous behavior, the presence of several species in the area and the wide range of wild and crop hosts of TSWV (Cho et al. 1989) complicates the development of pest management programs. The high cost of production for fresh market tomatoes reduces the options of pest management tactics and results in chemical control used in a preventive manner being used as the predominant tactic to ensure quality of the fruit.

As a new problem in the area, several characteristics of the vector-virus relationship need to be understood in order








5

to develop an appropriate pest management strategy. A sampling program suitable for management and research purposes is a necessary first step to understanding the population dynamics and other biological characteristics of this complex problem.

This study was oriented to

1. Determine the species composition and their seasonal

abundance and dispersion characteristics of the most important thrips species inhabiting tomato flowers

in North Florida.

2. Develop a sampling program for thrips on tomato

suitable for scouting and research programs.

3. Characterize damage to tomato fruit by Frankliniella

spp. thrips.

4. Determine the relationship between density of

Frankliniella spp. thrips and tomato fruit injury.








6




CHAPTER 2

LITERATURE REVIEW

Insects and Related Pests in Tomato

Tomatoes can be grown for processing or for fresh market, the latter being the most important in Florida (Clough 1987). Their production methods differ greatly; however, the range of potential pests is similar in all tomato crops grown in the same area and season under similar conditions.

Tomatoes are hosts for many kinds of arthropods whose importance varies within and among major production areas. Even when a few species are considered major key pests, there are many minor or secondary pests that under specific situations can become major pests. Their damage can cause an economic decrease in optimum yield or economic loss due to reduction in quality of the fruit below fresh market grade or for processing. The most common pests that directly destroy or affect the marketable product in Florida are the corn earworm, Heliothis zea Boddie; the southern armyworm, Soodoptera eridania (Cramer); and the tomato pinworm, Keiferia lvcopersicella (Walsisngham). Pests that reduce yield are the granulate cutworm, Feltia subterranea (Fab.); the serpentine leaf miner complex, Liriomyza spp.; and the tobacco hornworm, Manduca sexta (Joh.). Additionally, other pests are important in other areas or can become important in Florida: stink bugs,








7

lygus bugs, thrips, mole crickets, flea beetles, whiteflies, aphids and mites. Aphids, leafhoppers, threehoppers, thrips, and whiteflies may transmit viruses or mycoplasmas to tomato (Cantelo and Webb 1980, Lange and Bronson 1981, Pena 1983, Pohronezny et al. 1986).

Predators and parasites that attack pest species are frequently abundant, maintaining certain pests at low population levels. Among the most conspicuous are lady beetles, bigeyed bugs, Geocoris spp., minute pirate bugs, Orius spp. and parasitic Hymenoptera. Unnecessary application of pesticides can affect these natural enemies and thereby contribute to outbreaks of secondary pests.

Present tomato pest management systems utilize many resources, including host plant resistance, cultural control, natural and applied biological control, and chemical control. This is especially true for the most important pests. However, chemicals are widely used in a preventive manner to control them and ensure quality of the product. The use of pesticides varies from year to year and from one area to another, and may be limited by residue tolerance regulations (Tingey 1979, Lange and Bronson 1981).

Lange and Bronson (1981) consider that there is a definite trend away from the constant use of pesticides on tomato and recommend their use only when needed but in the context of a well-rounded holistic or integrated pest management approach to insect suppression. Research is needed








8

on sampling methods for some of the important tomato pests and most of the secondary pests. Establishment of economic thresholds and economic injury levels are also needed. These areas of investigation are difficult because of differences in cultivars, geographic and climatic areas, use of the crop, and many other variables.

Flower thriDs Frankliniella sn

General Structure and Systematics of the Order ThvsanoDtera

Thrips are minute and slender-bodied insects (0.5 - 14 mm in length) that may easily pass unnoticed. Their most striking feature is the wide marginal fringes on the long and slender, usually present, four wings which gives the order its scientific name. Nearly all true leaf and flower thrips have fully developed wings (macropterous) but apterous forms (brachypterous) are common in one or both sexes of many species as found by Eddy and Livingstone (1931) in Ef. fusca. The head bears a pair of usually 7 or 8-segmented antennae inserted at the front between the compound eyes. Three ocelli are also present on top of the head of most adults. The mouth parts, considered a unique characteristic feature in thrips, are of the sucking type with a stout, conical, and assymetrical proboscis (mouth cone) protruding beneath the head and often appearing to originate near or between the base of the front legs. The labrum and clypeus form the front of the proboscis, the basal portions of the maxillae form the sides, and the labium forms the rear. Three flexible stylets,








9

the left mandible (the right is rudimentary) and two maxillary stylets, lie within this cone. The first thoracic segment is freely movable (Lewis 1973, Borror et al. 1976).

The order Thysanoptera is divided into two suborders: Terebrantia and Tubulifera. Terebrantia has the last abdominal segment more or less conical or rounded, the females usually have a well developed saw-like ovipositor, the wings lie parallel to each other, and the males are always smaller and usually paler in color than females. Tubulifera, on the other hand, has the last abdominal segment drawn out into a tube, the wings overlap at rest so that only one is completely visible, the females lack an ovipositor and the males are often stouter than females with enlarged forelegs (Lewis 1973, Borror et al. 1976, Ananthakrishnan 1979).

The classification of families in Thysanoptera remains controversial among specialists; however, Priesner's scheme (cited by Stannard 1968, Lewis 1973, and Ananthakrishnan 1979) has been widely accepted. Priesner's classification involves five families, one of them (Phlaeothripidae) in the suborder Tubulifera, and the others (Aelothripidae, Merothripidae, Heterothripidae, and Thripidae) in Terebrantia. The great majority of thrips including most of the species of economic importance belong to the family Thripidae, which is widely distributed throughout the world.

Frankliniella, included in the family Thripidae, is a widespread genus through the world containing more than one








10
hundred, mostly flower living species (Palmer et al. 1989). Karny, cited by Sakimura and O'Neill (1979), erected the genus Frankliniella in 1910 and published the first key to world species two years later. Bailey (1957) considers that this genus is one of the most difficult to classify in Thysanoptera. According to the same author nearly 150 species have been described or transferred in and out of the genus. For example E. occidentalis, E. fusca, and E. tritici were first included in the genus Euthrips (Morgan 1913). Bailey (1940) considered Frankliniella spp. to be among the seven most injurious thrips groups of major importance in the United States where the genus includes the following species: E. bispinosa (Morgan), which is limited largely to Florida; E. fusca Hinds, found widespread throughout the continent; E. aosevpiana Hood restricted at that date to Arizona and California; E. insularis (Franklin), known from Florida, Texas and Arizona; E. moultoni Hood and E. occidentalis (Pergande), found commonly in the far western states but reported in Florida in 1910 by Morgan (1913); E. tritici (Fitch), abundant in the central and eastern states; E. vaccinii Morgan, limited to Maine and E. williamsi Hood, known from South Carolina and Florida. The apparent absence of some Frankliniella species in many states may be due to the lack of published data and the scarcity of collected material from these areas at that date.








11

Even when a knowledge of the species present may be of no significant importance because control measures for one will be equally effective against others, under some specific circumstances it may be highly desirable to recognize the species involved. It is especially true if we want to correctly apply an integrated pest management approach and we know that usually there are important biological differences to consider. For example, E. fusca requires approximately five days longer to complete its development than does E. tritici (Watts 1936). E. occidentalis and E. fusca are vectors of tomato spotted wilt virus, while E. tritici is not (Sakimura 1953, 1962, 1963).

Identification of thrips at the generic and specific levels is very difficult and usually requires examination of specimens properly mounted on a glass slide with a compound microscope. Keys for species identification in the genus Frankliniella have been developed by Watson (1923), Stannard (1968), and more recently by the Thysanoptera specialist R. Beshear (unpublished data) (Table 2.1 and Figure 2.1). The criteria most used by thysanopterists in Frankliniella include number, shape, and size of antennal segments and setal type of antennal sensoria, etc. (Kono and Papp 1977).








12




Table 2.1. Key to the females of Frankliniella spp.
thrips found in tomato flowers.*


1. Macropterous; abdomen either pale yellow with brown

patches on the meson or entirely dark brown .......... 2


Brachypterous; color generally dark brown to lighter

brown especially the thorax and head (in part) ... fusca


2. Pedicel of antennal segment III with a distinctly

thickened middle ring wich in profile appears as

angulations; setal comb on posterior margin of

abdominal tergite VIII incomplete .................... 4


Pedicel of antennal segment III straight or nealy

straight along sides; setal comb on posterior margin

of abdominal tergite VIII complete or incomplete ..... 3


3. Setal comb on posterior margin of abdominal tergite

VIII complete; anteromarginal and anteroangular setae

on pronotum of similar length; postocular setae longer,

as stout as the interocellar setae (Fig. 2.1).

*.......*............................. occidentalis








13



Table 2.1 (cont.)

Setal comb on posterior margin of abdominal tergite VIII incomplete; anteroangular setae usually longer

than anteromarginal setae on pronotum; postocular

setae shorter and much more slender than the interocellar pair (Fig. 2.1) ........................... fusca



4. Pedicel of 3rd antennal segment with distal and

basal parts equally diverging (Fig. 2.1) ........ tritici



Pedicel of 3rd antennal segment strongly angulated

12-14 um wide with the distal and basal parts of the angulation diverging inward; anteromarginal setae on

the pronotum usually less than 2/3 the length of

anteroangular setae (Fig. 2.1) ................ bispinos



Pedicel of 3rd antennal segment has smaller angulation

usually 7.0-10.5 um wide, the distal surface of the

angulation is flat but the basal part diverging:

anteromarginal setae usually longer, 0.7-0.8 times

longer than the anteroangular setae (from Miami area)

............................................... Ce halica



* With permission from R. Beshear (unpublished data)






F.occidentalis F.tritici F. fusca F. bispinosa ANTENNAE
segments l-III




HEAD




PROTHORAX fly' * ,'
(pronotum)




ABDOMEN
tergites VIII-X




Fig. 2.1. Key to the adult females of Frankliniella spp. thrips found in tomato flowers.








15


Biolovy of Frankliniella sp.

Sexual reproduction is prevalent among the Thysanoptera; however, parthenogenesis occurs in many species. Females often predominate in field populations and in some species males are rare or unknown (Lewis 1973). Watts (1936) found that E. tritici females greatly outnumber the males in nature from early spring through most of the summer, but during the autumn the proportionate difference is not so great. The same author also noted that all the progeny from unfertilized females were males. This last observation was also made by Eddy and Livingstone (1931) and Bryan and Smith (1956) for E. fusca and E. occidentalis, respectively.

The metamorphosis of thrips is intermediate between simple and complete. It is considered simple because the immature stages are very similar to the adults and more than one preadult instar has external wings. This metamorphosis is also considered complete because at least some of the wing development occurs internally and there is a quiescent instar preceding the adult. In Frankliniella spp., as most of the Terebrantia, the third and fourth instars are inactive, do not feed, and have external wings (Borror et al. 1976). Because of the confusion surrounding the type of metamorphosis of this order, variations in terminology for the immature stages have resulted and are common in the literature. Customarily, thysanopterists have called the first two feeding instars








16

larvae and the remaining quiescent, nonfeeding, preadult instar prepupa and pupa (Watts 1936, Stannard 1968, Lewis 1973, Nugaliyadde and Heinrichs 1984). These traditional terms have been objected to by some entomologists who have supported the use of the terms nymph and pseudopupae (Bryan and Smith 1956, Bailey 1957, Lublinkhof and Foster 1977, Ananthakrishnan 1979).

Frankliniella spp., like other genera in the family Thripidae, are oviparous. E. occidentalis females begin oviposition within 72 hours after adult emergence and oviposition normally continues intermittently throughout remaining adult life (Bryan and Smith 1956). Watts (1936) observed that during the colder months the oviposition period can be drastically reduced in E. tritici. The total number of eggs laid by most female thrips ranges from about 30 to 300 (Lewis 1973). Robb et al. (1987) found that an adult E. occidentalis can deposit 150-300 eggs during its life. Watts (1934) reports an average of 55.5 and 41.5 eggs per female in E. fusca and E. tritici, respectively. The same author (Watts 1936) found later that E. tritici oviposited an average of 28.61 eggs per female with a maximum of 119. This large variation in the number of eggs laid by different females suggests the possibility that a larger average could occur in the field under favorable conditions. Lublinkhof and Foster (1977) showed that temperature can dramatically affect the reproductive rate of �. occidentalis females. They found a








17

mean of 24, 95 and 44 offspring per female at temperatures of 15, 20 and 300C, respectively.

Most Terebrantia deposit their eggs singly in an incision made in the plant tissue by means of the saw-like ovipositor. Oviposition can occur within leaf sheaths, at the flower base, cotyledons, petals, sepals or glumes (Lewis 1973). Seedling cotton was successfully used as oviposition media when rearing E. tritici (Watts 1936). Bryan and Smith (1956) used bean pods for oviposition of E. occidentalis. This same species is known to oviposit in the young berry grapes (Yokoyama 1977). Peanut leaves have been used as oviposition and rearing media of E. fusca (Kinzer 1968, Kinzer et al. 1972).

The duration of the life cycle varies with the species and temperature (Table 2.2) Watts (1934) found that E. fusca has a longer life cycle and lives longer in the adult stage than E. tritici under identical conditions. A similar duration of the life cycle of E. fusca at 270 C is reported by Kinzer et al. (1972). Lublinkhof and Foster (1977) report similar data for E. occidentalis. These authors also found a significant effect of temperature on developmental time. All life stages developed more rapidly as temperatures increased.

Ecoloy of Frankliniella spp,

Host-plant relationshiDs

Flower thrips are very general feeders and are capable of reproducing on a wide range of plant species. Lists of







Table 2.2. Developmental rate of Frankliniella spp.



Temp. Larvae Pre Egg to Adult

Species oC Egg I II pupa Pupa adult longevity Source



F. occidentalis 15 11.2 4.9 9.1 2.9 5.6 33.7 70.8 20 6.4 2.3 5.2 2.2 2.8 18.9 56.7 30 4.3 1.1 4.3 1.4 1.6 12.7 27.5 (1)



F. tritici 3.3 2.0 2.4 1.1 2.3 11.1 21.4 (2)



F. fusca 6.7 2.3 3.5 1.1 2.6 16.3 29.1 (2) 27 -- -- -- -- -- 16.0 -- (3)



1. Lublinkhof and Foster (1977)

2. Watts (1934)

3. Kinzer et al. (1972)








19

recorded hosts have been reported for E. occidentalis (Bryan and Smith 1956, Yudin et al. 1986, Stewart et al. 1989), E. tritici (Watts 1936, Stewart et al. 1989), and E. fusca (Eddy and Livingstone 1931, Newsom et al. 1953). These lists include abundant and diverse host plant species belonging to several unrelated families and orders. Annual and perennial weeds as well as winter, summer and perennial crops, recorded as hosts, provide the flower thrips with an assortment of uninterrupted suitable hosts during the year. Movement. migration, and dispersal
Thrips are considered weak flying insects even though flight is by far their most important natural form of dispersal. Regular dispersal can occur by self-directed flight or by passive movement by wind currents. Their finely fringed wings enable them to remain airborne long enough for the wind to blow them to great heights and for long distances (Stannard 1968, Lewis 1973). Despite their great dependence on wind for dispersal some species seem to be able to see, orientate towards, and land on a surface of their choice. Yudin et al. (1987) found that E. occidentalis was highly attracted to white traps. Some species of Frankliniella are more attracted to some plant species, varieties, or cultivars (Watts 1936, Yudin et al. 1988, Stewart et al. 1989).

Mass flights occur sporadically within the flight period and in many species are reputedly associated with thundery weather (Lewis 1973). Stannard (1968) suggested that








20

migration of F. tritici over large continental land masses might be associated with frontal winds. Survival and natural regulation

Overwintering of Frankliniella spp.has not been studied in detail. Lewis (1973) considers that only a few species of thrips do not overwinter. Watts (1936) found that winter conditions in South Carolina permit continuous development of immature stages of E. tritici, though in smaller numbers and at a slower rate than during the summer. He found second instar larvae on oats shortly after a low temperature of 12oF. E. occidentalis has also been found during the winter in California (Bryan and Smith 1956) and southern Texas (Stewart et al. 1989). Eddy and Livingstone (1931) reported that only E. fusca females survive the winter in South Carolina.

Temperature and rainfall are the two most important weather factors affecting the number of thrips (Lewis 1973). Bryan and Smith (1956) considered that the distribution of rainfall may be more important than the total amount. Heavy rains of short duration probably limit the increase in population by beating large numbers of adults and immature forms into the ground. Similar observations are made by Watts (1936). However, there are insufficient data to support these observations.

Besides hazards of climate and change in the environment, many predators and parasites regulate thrips populations. Thrips are eaten by many general predatory insect species in








21'

the orders Hymenoptera, Diptera, Coleoptera, Hemiptera, Neuroptera, and even in Thysanoptera. Mites in the families Laelaptidae, Cheyletidae, and Amystidae have also been reported attacking thrips. Additionally, thrips form part of the diet of some amphibians, reptiles and insectivorous birds or can be eaten accidentally. Despite their small size, the eggs and larvae of thrips are parasitized by minute wasps in the families Eulophidae, Trichogrammatidae, Mymaridae and Scelionidae. Nematodes infesting the larvae, pupa, and adult thrips have also been reported. Mites are also known to be ectoparasites of thrips. Fungi belonging to the genera Beauvaria, CeDhalosporium, and Entomopthora also induce mortality in natural populations (Stannard 1968, Lewis 1973, Ananthakrishnan 1979). This long list of natural enemies of thrips suggests the potential possibilities of biocontrol agents. However, few attempts to introduce or encourage natural enemies to control pest thrips species have been reported.

Economic Importance

As pests on several crops

Several hundred species of thrips are considered pests in many crops. They can cause serious direct damage to crops by feeding, oviposition or as vectors of plant pathogens. Direct injury produced is often slight but occasionally it may be severe and result in serious losses when heavy infestations occur. Thrips feed on most plant parts, except roots, but








22
feeding is usually concentrated on rapidly growing tissues such as young leaves, flowers and terminal buds. The typical injury to plant tissue consists of silvering, scarring and distortion of leaves and fruits caused by feeding, and subsequent later discoloration due to excrement and the growth of molds (Lewis 1973). Oviposition in newly developing fruits can produce a halo spot consisting of a small dark scar surrounded by whitish tissue (Jensen 1973). Beside causing mechanical injury, some thrips transmit plant pathogens including toxaemias caused by toxin in salivary secretions, bacterial and fungal diseases spread by mechanical contact and viruses transmitted during feeding (Stannard 1968, Lewis 1973, Ananthakrishnan 1980, Walkey 1985).

Bailey (1940) listed the 32 thrips groups most destructive to crops in the United States. He placed the Frankliniella group among 7 groups or species of major importance. E. occidentalis is considered an important pest in cotton (Stoltz and Stern 1978a, b, Rummel and Quisenberry 1979, Wilson 1982, Graves et al. 1987), table grapes (Jensen 1973, Jensen and Luvisi 1973, Stafford 1974, Yokoyama 1977), apples (Terry and DeGrandi-Hoffman 1988), safflower (Carlson 1964b, 1966, Carlson and Witt 1977), onions (Harding 1961, Carlson 1964a, Dintenfass et al. 1987), tomato (Oetting 1985, Olson and Funderburk 1986, Schuster and Price 1987, Cho et al. 1989), ornamentals (Baker and Stephan 1986, Zur Strassen 1986), lettuce (Yudin et al. 1987, 1988), and, because of its








23

extensive host range, probably. in other crops not yet reported.

E* fusca has been reported attacking peanuts (Morgan et al. 1970, Tappan and Gorbet 1979, 1981 Lynch et al. 1984, Tappan 1986a, b) and cotton (Watts 1934, 1937a, b, Watson 1965, Rummel and Quisenberry 1979, Eddy and Livingstone 1931). However, its status as an economic pest is not clear. Watson (1965), studying the control of thrips (including E. fusca and E. tritici) in seedling cotton, found that where thrips were sufficiently abundant to produce plant damage, thrips control resulted in more uniform and vigorous plants but did not influence the rate of crop maturity or the total yield of seed cotton. Similar results have been obtained in peanuts by Morgan et al. (1970). Additionally, E. occidentalis and E. fusca are vectors for the tomato spotted wilt virus (Sakimura 1962, 1963, Paliwal 1976, 1979, Cho et al. 1984, 1989, Allen and Broadbent 1986, Greenough et al. 1985).

E. tritici has been reported in soybean (Marston et al. 1979, Irwin and Yeargan 1980), cotton (Watts 1934, 1936, Watson 1965, Rummel and Quisenberry 1979), peanuts (Morgan et al. 1970), and strawberry, roses, alfalfa, peas, peach, apricot, plum, etc., which are damaged to varying extent by this species (Watts 1936). E. bispinosa has been reported in peanuts (Morgan et al. 1970).








24

As predators of other pest species

Predation on other small animals is a common phenomenon among thrips. Some species are entirely carnivorous, however, most predatory thrips are usually polyphagous including animal and plant materials in their diet. Predatory thrips species feed on small, soft-bodied insects, including other thrips, aphid nymphs, psocids, scales, and especially on the eggs of mites and Lepidoptera (Lewis 1973). Probably the best documented case of predation of the genus Frankliniella is in cotton and peanuts. E. occidentalis is an omnivore, consequently an important early-season predator of Tetranychus spp., an important pest in cotton (Trichilo and Leigh 1986, Pickett et al. 1988, Gonzalez et al. 1982). Boykin et al. (1984) listed E. fusca and E. tritici as predators on eggs of the two spotted spider mite, T. urticae, Koch on peanut fields. E. occidentalis is considered by Lewis (1973) an important predator on Aelothrips fasciatus and Tetranychus telarius (L.), and E. fusca on 2. telarius. Despite the apparent effectiveness of Frankliniella spp. as predators on mites, its potential as a biological control agent may be limited because of its important potential as a pest of many crops.

Pollination of flowers

Another beneficial characteristic of thrips is their role in the pollination of flowers. Their individual contribution is probably small but compensated for by the great numbers








25

that may be present in flowers. For example, F. occidentalis has been reported to increase pollination on onions (Carlson 1964a). Individual E. tritici captured while flying carried large quantities of pollen among flowers, thereby playing an important role on cross-pollination of several plant species (Lewis 1973).

The economic importance of Frankliniella spp. as predators and/or pollinators needs to be better analyzed and is probably insignificant. However, these characteristics must be considered when implementing management programs in an IPM approach.

Tomato Spotted Wilt Virus (TSWV)

Tomato spotted wilt virus is a serious disease affecting the production of many food and ornamental crops worldwide (Greenough et al. 1985, Allen and Broadbent 1986, McRitchie 1986, Cho et al. 1986, 1989). TSWV was first described in Australia by Brittlebank in 1919 (cited by Cho et al. 1989). It has since become an important problem in the United States and Canada (Cho et al. 1988).

TSWV is unique among plant viruses in that it is the only member of its group for which identification is based on particle morphology, has one of the broadest host ranges of any plant virus, is highly unstable in vitro, and is the only

virus transmitted in a persistent manner by certain species of thrips (McRitchie 1986, Cho et al. 1989). Their large isometric particles are 85 nm in diameter and are enclosed








26

within a lipoprotein envelop constituting about 20% by weight of virus particles (Bos 1983, Walkey 1985). Other chemical and physical characteristics of TSWV may be found in Ie (1970). Host Rane of TSWV

TSWV, like its vectors, has a very wide host range that includes at least 200 plant species. While dicotyledonous plants (at least 192 species in 33 families) seem to be preferred, monocotyledonous species may also be infected (8 species in 5 families) (Cho et al. 1989). Economically important crops affected include tomato, potato, bell pepper, tobacco, lettuce, peanut, pineapple, papaya, cucumber, and several flowering ornamentals such as chrysanthemum, dahlia, gloxinia, and gerbera daisies (Ie 1970, Baker and Stephan 1986, McRitchie 1986, Cho et al. 1989). Several annual and perennial weed hosts may serve as additional virus reservoirs. Cho et al. (1986) found 44 plant species (mostly weeds), representing 16 families, infected with TSWV in Hawaii. The wide host range of TSWV ensures its perpetuation and facilitates its spread into crop fields throughout the year in temperate and subtropical regions. This characteristic, on the other hand, complicates the implementation of strategies to manage TSWV.

TSWV Transmission

Transmission of TSWV can occur through seed, by sap (in laboratory), or by vectors. When transmission occurs through seed, the virus can be apparently carried in the teste but








27

not in the embryo. Infection by seed, however, is confused and rates of infection from one to 96% have been reported. Transmission by artificial inoculation of sap using abrasives is easy, especially when extracts are prepared in neutral buffer containing reducing agents (Ie 1970). Transmission by vectors is probably the most important means. Only thrips are known to be vectors of TSWV, and TSWV is the only confirmed virus transmitted by thrips (Ananthakrishnan 1980, Walkey 1985). Eight species of thrips belonging to the family Thripidae have been identified as vectors of TSWV: E. occidentalis, E. fusca, E. schultzei (Trybom), E. moultoni Hood, E. tenuicornis (Uzel), Thrips tabaci Lindeman, 2. setosus Moultan, and Scirtothrips dorsalis Hood (Sakimura 1953, 1962, 1963, Paliwal 1979, Ananthakrishnan 1980, Amin et al. 1981, Allen and Broadbent 1986, McRitchie 1986, Cho et al. 1988, 1989). However, Paliwal (1976) found that 2. tabaci did not transmit the virus in Canada.

TSWV is transmitted by thrips in a persistent and circulative manner. Viruses transmitted in this way are characterized by a long acquisition feeding time, a latent period (12 hours or more) after feeding before transmission can occur, vectors retain the ability to transmit it for at least a week or for the remainder of their life, and finally the virus is retained through the moult of the insect (transstadial transmission) (Walkey 1985). TSWV is acquired only in the larval stage of thrips but may be transmitted








28

throughout the adult stage. The acquisition feeding period is 2-3 days for 2. tabaci, E. occidentalis, and F. fusca (Sakimura 1962, 1963) with a shortest reported period of 15 minutes for 2. tabaci (Ie 1970).

A 4-to 10-day incubation (latent) period, depending on the vector species, occurs before the virus can be transmitted (Ie 1970). Sakimura (1963) found that the average latent period for E. fusca was 9.3 days with a range of 4-12 days. Transmission efficiency was increased with acquisition feeding times from fifteen minutes to four days (Walkey 1985). Vectors are maximally infective 22-30 days after acquisition but sometimes retain the virus for life (Ie 1970). The general average in retention period observed in E. fusca and 2. tabaci was 15.4 and 11.9 days, respectively, with a wider range in E. fusca (1-43 days) (Sakimura 1963). Transovarian passage of TSWV has not been reported, and it is speculated that vectors do not transmit the virus to their progeny (Ie 1970, Walkey 1985). Males and females and macropterous and brachypterous forms of E. fusca did not differ significantly in their vectoring ability (Paliwal 1976). Enzyme-linked immunosorbent assay (ELISA) can be used to detect TSWV in individual thrips (Cho et al. 1988).

SvmDtoms of TSWV

Characterization of the symptomatology of TSWV is difficult because of its wide host range. Hosts may show differing symptoms because of specific responses of the








29

affected plant species and in part due to the many strains of the virus (McRitchie 1986). Many minor variants (strains)have been isolated which produce symptoms differing in severity. The most stable and important of these are: strains TB (tip blight), N (necrotic), R (ring spot), M (mild), VM (very mild); strains A, B, CI, C2, D, and E; the "vira-cabeca", and the tomato tip blight strain (le 1970).

Symptoms of TSWV can include local necrotic lesions followed by systemic necrotic patterns, leaf deformation, local black spots, systemic mosaic pattern of yellow and dark green specks, local chlorotic spots with necrotic centers, etc. The list of symptoms goes on and on and can be easily confused with symptoms induced by other viruses, fungal or bacterial pathogens, or nutritional disorders (Ie 1970).

Symptoms on foliage of infected tomato plants are characterized by thickening of veins, downward curling of leaves, bronzing, and ring spotting. The entire plant is often stunted. However, symptoms are most noticeable and diagnostic on immature tomato fruit. Green fruits show light green rings with raised centers, giving the fruit a lumpy appearance, and rendering them unmarketable (McRitchie 1986). Varied and complex symptoms patterns can be found in ripe tomato going from almost unnoticeable decoloration of the normal red color of the fruit to very well marked mosaic of different tones of red, yellow, and green (personal observation).








30

Management of Frankliniella SDD. and TSWV Pest Management in Tomatoes

Pathogens, insects, nematodes, and weeds are of constant concern to tomato growers and in many cases limit production. These pests are especially important for fresh market tomato because high quality of the fruit is required. Tomato growers in Florida have adopted a complex horticultural scheme to deal with these pests. Tomatoes are grown on raised, plasticmulched beds fumigated with broad-spectrum biocides before planting. Pesticides have been traditionally sprayed every 37 days to prevent foliar pest problems. Frequently, a single problem can require special treatment. Approximately 34 insecticide applications have been made on a single crop in an attempt to control leafminers (Liriomyza spp.) in South Florida. Up to 60 fungicide sprays have been applied on some winter crops to prevent diseases, especially bacterial spot (Xanthomonas campestris pv. vesicatoria) and late blight (Phytophtora infestans (Mont.)), and still heavy losses have sometimes resulted. These examples give an idea of the amount of money spent by tomato growers when dealing with pest problems. Approximately 20% of the money invested in tomato production is spent for pesticide spray materials (Pohronezny et al. 1986).

Despite the predominance of chemical control being widely used in tomato pest management, other tactics have been developed. Lange and Bronson (1981) consider that, beside








31

pesticides, present tomato pest management systems utilize many resources including host plant resistance, cultural, and biological controls. However, the same authors stated that many kinds of constraints against development and utilization of pest management programs exist, including fruit quality standards, pesticide use limitations, and research on improving management programs.

Several adverse consequences have been associated with the constant use of pesticides, such as increased costs of production, development of resistance to the insecticides used, reduction of predator and parasite populations, difficulty in registering new products, environmental contamination, and social concerns. These factors have created a definite trend away from the constant use of pesticides in the most important tomato production areas in the United States (Lange and Bronson 1981, Anonymous 1985, Pohronezny et al. 1986).

The complex nature of the tomato pests requires that researchers from different disciplines participate in the development and implementation of management methods. IPM programs, oriented primarily to use pesticides only when needed, are being used in many tomato producing areas such as Florida (Pohronezny et al. 1986), California (Anonymous 1985), and other states and countries around the world (Lange and Bronson 1981).








32

IPM, as defined by Bottrell (1979), is the selection, integration, and implementation of pest control (tactics) based on predicted economic, ecological, and sociological consequences. Under this approach, pests are treated as part of a crop production system that includes also the physical and biological environment. The emphasis of IPM programs is on anticipating and preventing problems when possible. Four components are essential in any IPM program: accurate identification of pests, field monitoring, control action guidelines, and effective methods for prevention and control (including the correct use of appropriate pesticides when needed) (Anonymous 1985). Some of these components such as pest identification and control methods are self explanatory. Monitoring, by routine field checks, provides the information necessary to evaluate pest problems and make management decisions. Control action guidelines are numerical thresholds which reflect the population level that will cause economic damage and provide a way to decide whether actions are needed to avoid eventual losses.

Monitoring was considered by Pohronezny et al. (1986) the most important feature of a tomato IPM program started in Florida in 1976. Under this program, fields were systematically scouted twice-weekly and the resultant biological data used in farm management decisions. The basis for monitoring are given by the development of sampling programs which include several components: sampling technique,








33

number of samples, sample unit, factors affecting density estimates, sample allocation, spatial and seasonal distribution, insect stage to sample, etc. (Ruesink . 1980, Wilson 1985, Pedigo 1989). These components will be discussed in section 4.3.

Despite the great interest in implementing IPM programs to control pests in tomato, Lange and Bronson (1981) stated that the present state of this technique is just at the threshold of understanding the intricacies of this pest-crop complex which also includes economic and social characteristics. Differences in cultivars, geographical areas, use of the crop, costs of production, cohsumer requirements on quality, etc. are some of the limitations to the implementation of IPM programs in tomato. Control of Frankliniella sop. and TSWV

The tactics that can be used effectively in the control of plant virus diseases involve evasive measures by reducing sources of infection, limiting spread by vectors, and to minimize - the effects of infection on yield. Sources of infection can be reduced by destruction of infected plants in the crop, in other crop species, or weeds known to be reservoir hosts of the virus; planting virus-free seeds or vegetative stocks, modifying planting and harvesting procedures and dates (breaking an infection cycle). Control of vectors is probably the procedure most commonly used to avoid the disease. However, breeding for resistance or








34

immunity to a virus provides the best single solution to the

problem of virus disease (Matthews 1970). Naturally, the combination of two or more of the measures above described will improve the prevention of virus diseases. It is also essential that all growers follow those measures because if one fails in doing it, more sources of infection may be available.

Probably the most complete research program oriented to the integrated management of TSWV is the one being conducted in Hawaii by Cho et al. (1989). These scientists identified some useful tactics that combined can minimize TSWV disease occurrence. These tactics include crop rotation with nonsusceptible crops to reduce the buildup of inoculum sources, crop placement to avoid planting TSWV-susceptible crops adjacent to each other, control of alternate TSWV vector hosts, use of virus-free seedling, regular applications of insecticides, fallow field areas with high disease incidence, and soil fumigation to eliminate thrips associated with crop debris. However, their conclusions stated that these tactics are not totally effective if virus and vector occurrence is high and. frequent throughout the area. Under these conditions they suggest not to continue planting susceptible crops. In another area of research, the same authors are developing resistant varieties of tomato, having developed at present several lines highly resistant to TSWV.








35

The only available- method to control TSWV in tomato is through the use of insecticides to reduce the thrips vectors. Control of adult thrips moving into tomato fields is effective in reducing primary spread of TSWV and direct cosmetic damage to the fruit due to oviposition, while control of immatures reduces secondary spread of the disease.

Many pesticides have been shown experimentally to effectively control Frankliniella spp. populations in several crops. These insecticides include abamectin, acephate, aldicarb, azinphos-methyl, carbofuran, chlorpyrifos, cyfluthrin, cypermethrin, diazinon, endosulfan, fluvalinate, formetanate hydrochloride, metamidophos, methomyl, mevinphos, monocrotophos, naled, phorate, and phosphamidon (Carlson 1964a, Morgan et al. 1970, Carlson and Witt 1977, Tappan and Gorbet 1981, Lynch et al. 1984, Oetting 1985, Robb et al. 1987, Cho et al. 1989). This long list suggests that chemical control of Frankliniella spp. is feasible and resistance to insecticides can be avoided or minimized by properly rotating chemicals. However, chemical control of these species seem to be uneconomical in some crops. For example, despite effective reduction of Frankliniella spp. populations in peanuts, onions, and cotton, yields were not increased when control measures were used (Carlson 1964a, Watson 1965, Morgan et al. 1970, Tappan and Gorbet 1981, Lynch et al. 1984). Similar results have been observed when trying to control TSWV or cosmetic damage due to F. occidentalis oviposition. Cho et al.








36
(1989) found that none of six insecticides, which were effective to control the vector, suppressed TSWV disease occurrence in lettuce. Metamidophos has proven effective in reducing Frankliniella spp. populations in tomato fields in North Florida, but cosmetic damage to the fruit has remained high even when the frequency of pesticide application has been increased (personal observation), probably because invading populations reinfest the field.

The economics of chemical control of Frankliniella spp. in some crops such as cotton, peanuts, and onions remains in a state of confusion and most researchers agree that control with pesticides is uneconomical. The situation in tomato is different; its production could be probably impossible without the use of insecticides if TSWV and its vectors, which can additionally cause cosmetic damage, are present.

Some herbicides have been reported to have certain insecticidal properties on thrips. Laster et al. .(1984) found that the herbicides dinoseb and dinoseb + MSMA showed some insecticidal activity against Frankliniella spp. when applied as a standard weed control in cotton. Dinoseb has been discontinued but similar possibilities could be considered.

Biological control is another alternative for managing Frankliniella spp. Predators and parasitoids usually kill large proportions of field populations of thrips (Lewis 1973); unfortunately, the constant use of insecticides reduces their densities. Stoltz and Stern (1978a) found that reduction of








37

Geocoris pallens Stall from the multiple dimethoate and naledtoxaphene applications to control lygus bugs, allowed populations of its prey E. occidentalis to increase significantly.

Other predators on F. occidentalis are the minute pirate bug, Orius tristicolor White (Stoltz and Stern 1978b) and Q. insidious (Say) (Lewis 1973). Other predators, parasites, and pathogens inducing mortality in natural populations of thrips are potential sources of biocontrol agents for Frankliniella spp. and are mentioned by. Stannard (1968) , Lewis (1973), Ananthakrishnan (1979), and Young and Welbourn (1988). However, few attempts (if any) to introduce or encourage natural enemies to control these pest species have been reported.

The development of resistant varieties to thrips damage could be another alternative of control. Unfortunately, breeding for resistance to these pests has received very little emphasis, probably because resistant varieties are considered of not much value for a cosmetic damage problem like that produced by thrips on tomato fruit. However, resistant varieties of table grapes to F. occidentalis damage (similar to the damage found in tomato) were reported by Jensen (1973). He found that while some varieties are not affected, the greatest damage occurs with the variety "Italia" whose skin may split at the site of a halo spot during ripening and lead to bunch rot. Other examples of resistant








38

varieties to Frankliniella spp. occur on peanuts, where the Spanish, Valencia, and Virginia type peanut plants are resistant to �. fusca (Smith 1979), and cotton, where glabrous genotypes showed some resistance to thrips (Rummel and Quisenberry 1979).

Insecticides seem to be an obligatory alternative tactic when economic loss due to reduced quality of fruit for fresh market or for processing can be expected, or if TSWV is present in the area. This is the case with Frankliniella spp. in tomato in North Florida. Under these conditions, the immediate focus of pest management has to be directed to develop a program for the more precise and judicious application of pesticides, while other alternatives are evaluated and implemented.

Surveillance and Sampling Programs

Acquiring quantitative and qualitative information about the agroecosystem is a preliminary and indispensable phase in IPM programs. Any action to be taken against a pest will depend on its presence, abundance, and potential threat, documented through pest surveillance. Pedigo.(1989) defines pest surveillance as the watch kept on a pest for the purpose of decision making. The major objectives are detection of species present and determination of population density, dispersion, and dynamics. Population density estimates are obtained in quantitative surveys by sampling. Sampling requires that a representative part of the total population








39

be taken and estimates be based on that part. Therefore, surveillance requires the availability of sampling programs. The development of a sampling program requires an understanding of its components including spatial distribution, sampling technique (how to collect information for a single sample), location of samples, number of samples, when to start sampling, frequency of sampling, what life stage of the pest to sample, etc. (Wilson 1985, Pedigo 1989). Development of efficient, reliable sampling programs for insect pest management requires considerable research by entomologists in order to generate, quantify and understand biological information of the pest-host relationship and the factor affecting this interaction. Thris sampling techniques

Techniques of sampling thrips, that give qualitative and quantitative estimates of the size and distribution of populations, are briefly described in this review. A more complete description can be found in Lewis (1973), Southwood (1978), and Irwin and Yeargan (1980). Thrips can be sampled in three main environments; i.e., soil, vegetation, and air. The technique will vary for each habitat, for each species or group, and according to the objective of sampling. Populations in soil are usually taken with a 10-cm-diameter soil-corer. Litter can be sampled within a grid by scraping it to soil level and collected. Thrips can be extracted from samples of soil or litter by a dry funnel method in which the insects are








40

stimulated to move from samples by heat applied from above by light bulbs or infra-red heaters. This method is more efficient for adults than for larvae, and useless for pupae. Flotation techniques can also be used to separate specimens from soil and litter. Populations that emerge from soil in spring can also be sampled with emergence cages placed on the ground.

Thrips can be sampled from vegetation by removing infested specific parts of the plant (or the whole plant) and processing in the laboratory. Thrips can be separated from vegetation by washing, repellents, heated funnels, cooling, etc. Small structures of the plant (leaves, flowers, etc.) can be placed into 70% alcohol and thrips separated by hand. Estimates from standing vegetation can be taken with a sweep net, suction samples, beating and shaking, or directly counting the thrips on the plant part of interest.

Aerial population can be sampled with suction traps (absolute estimates) or sticky traps and water traps (relative estimates). Suction traps can be used to measure the daily periodicity of flight, the specific structure of faunas, or to relate numbers taking-off with weather conditions. Sticky traps are useful for detecting the presence of flying thrips and especially to record times of emergence and seasonal changes in activity. Some species of thrips are attracted to pale colors, especially white. The appropriate color for the species of interest must be determined, or black can be used








41

to avoid differential attraction when studying the species composition of an habitat.

Several of the above listed techniques have been used to estimate Frankliniella spp. populations in many crops. These can be divided into direct observation or collecting techniques and delayed collecting techniques. The direct count technique or visual inspection is that in which thrips are counted, from the plant or plant part of interest, directly in the field. This method was described efficient for Sericothrips variabilis (Beach) in soybean, but not for E. tritici (Irwin and Yeargan 1980). Letourneau and Altieri (1983) found the visual inspection technique more effective than sticky traps, pan traps, and malaise traps when sampling E. occidentalis in squash.

Direct collecting techniques are those in which thrips are collected without bringing plants or plant parts to the laboratory. These techniques include shaking or striking, aerial suction, D-vac suction sticky or cup traps, malaise traps, irritation, etc. Shaking has been used to sample E. occidentalis in apple (Terry and DeGrandi-Hoffman 1988) and grape blossoms (Jensen 1973) and Frankliniella spp. from whole cotton plants (Watson 1965, Laster et al. 1984). Aerial suction traps were found to be effective in sampling E. occidentalis in grapes (Yokoyama 1977). The D-Vac suction procedure has been used effectively for Frankliniella spp. in cotton (Stoltz and Stern 1978a, Marston et al. 1979).








42

Sticky traps have been used for sampling E. occidentalis in onions (Harding 1961) and E. tritici in roses and peonies (Webb et al. 1970). A modified sticky trap, using Styrofoamm or plastic cups, was used by Yudin et al. (1987) for E. occidentalis in lettuce. The same authors evaluated several colors and found that white traps caught significantly more thrips than 14 other colors tested including several tones of yellow, green, red, blue, etc. The cage-aerosol or irritation technique was used for sampling E. tritici in roses (Ota 1968) and in soybean (Marston et al. 1979). However, this procedure was less effective than other techniques in both cases.

Delayed collecting methods are those in which the whole plant or plant parts are removed for later processing in the laboratory to separate the thrips. Whole plants have been used for sampling E. occidentalis in onion (Dintenfass et al. 1987), cotton (Pickett et al. 1988), lettuce and weeds (Yudin et al. 1988), and several crop and non-crop plant species (Stewart et al. 1989). Flowers have been collected to sample Frankliniella spp. in apples (Terry and Degrandi-Hoffman 1988), roses (Henneberry et al. 1964, Ota 1968), peanuts (Tappan 1986a, b, Tappan and Gorbet 1979, 1981), tomato (unpublished data), and several weed species (Yudin et al. 1986). Terminal buds have been used to sample . tritici in soybean (Irwin et al. 1979), and E. fusca in peanuts (Tappan 1986a, b, Tappan and Gorbet 1979, 1981, Lynch et al. 1984). Leaves have been used in cotton (Pickett et al. 1988), soybean








43

(Irwin et al. 1979), weeds (Yudin et al. 1986), and pyrethrum (Bullock 1963).

Extraction of thrips from foliage is the next step in the delayed collecting technique. Several procedures can be used to separate thrips from foliage (Bullock 1963), including washing in alchohol or detergent (Ota 1968, Henneberry et al. 1964, Tappan 1986a, b, Tappan and Gorbet 1979, 1981, Dintenfass et al. 1987, Yudin et al. 1988) shaking devices (Henneberry et al. 1964), berlese funnels (Stoltz and Stern 1978a, Stewart et al. 1989), turpentine-vapor chamber (Yudin et al. 1986), and extraction of thrips by hand using an stereoscope (unpublished data). Spatial distribution Datterns

The determination of spatial distribution patterns is necessary in the development of sampling procedures for pest species and their natural enemies in crop systems. This structural component and description of the condition of a population, also called dispersion characteristics, refers to how a population occurs in space. Various spatial patterns, in space and/or time, can be displayed by insect populations, random, uniform, and clumped. These are described in Southwood (1978), Ruesink and Kogan (1975), Ruesink (1980), and Wilson (1985). The spatial distribution patterns of a species affects the sampling program and the method of transforming the data prior to statistical analysis. Most insects rarely disperse themselves randomly in their natural








44

environments. As a result their dispersion patterns do not follow a normal frequency distribution (Waters 1959). However, most statistical methods are applied to data with an assumption of a normal frequency distribution. In order to overcome these problems, the actual data has to be transformed by a function whose distribution normalizes the data or stabilizes the variance. Southwood (1978) recommends the transformation of data from a clumped population by using logarithms, slightly clumped by using square roots, and uniform by using squares.

A spatial pattern is random if every point on the surface has an equal probability of being occupied by an individual. In other words knowing the location of one other individual on the surface provides no information as to the location of any other individual. The poisson statistical distribution describes this spatial pattern. The clumped, aggregated, or contagious spatial pattern is described by the negative binomial distribution and occurs when the presence of an individual in a unit of habitat increases the probability of another occupying the same unit. This is the most frequent distribution encountered -in insects (Southwood 1978). Finally, the uniform or regular distribution occurs when the presence of an individual at one point decreases the probability of another being nearby.

A species distribution pattern can be expressed in terms of a variance:mean ratio (si). When s2/R=l (or s2=), the








45

distribution is considered random; if s2/i>1 the distribution is considered clumped; and if s2/i
An alternative method was proposed by Taylor (1961) to describe the relationship between sample mean and sample variance. Regardless of the organism considered or the sampling technique used, variance is considered to be related to the mean by the Taylor's Power Law: s2=ai~. Where A and B are species (and sample unit) specific coefficients which together describe, and can be used to classify a species distribution pattern. The coefficient A varies with sampling technique and habitat, while the exponent b is constant for the species. Coefficients A and B can be estimated by a log transformation of the variance - mean data, where b is the slope of the transformed data, and A equals the antilog of the transformed intercept. The most convenient method for estimating a and ] is via linear regression of log s2 on log


Many consider that Taylor's power law has provided a better model of variance/mean relationships than other models








46

(Elliot and Kieckhefer 1987, Funderburk and Mack 1989). It is also a useful tool in agricultural research for determining appropriate sample sizes for use in estimating the abundance of various species (Wilson 1985) and for developing binomial (or presence-absence) sampling programs (Wilson and Room 1983).

Number of samples

The total number of samples for estimating insect densities will depend on the degree of precision desired or required. Optimal sample size determination allocates the minimal number of samples needed to achieve a certain level of precision. The reliability of the estimated density increases when sample size is increased; unfortunately, the cost of sampling is also increased incrementally. A proper balance between the reliability of the estimate and the cost of obtaining it must be estimated (Ruesink 1980, Pedigo 1989).

The degree of precision required depends on the purposes of the sampling program. Ruesink and Kogan (1975) distinguish between data used for insect pest management and data for research purposes. In the first case the idea is to know if the population exceeds a given threshold and considerable sampling error can be tolerated especially if the sample mean is exceptionally far from the economic threshold. In the second case emphasis is placed on the reliability of the parameter estimate and the sampling error must be equally small for all sample means.








47

A series of equations for use in estimating sample sizes are described by several authors (Karadinos 1976, Southwood 1978, Ruesink 1980, Wilson .1985). Within a homogeneous habitat, Southwood (1978) considers that the number of samples

(n) required is given by: n=(s/ER) ; where: a represents the standard deviation, I the mean and X the -predetermined standard error as a decimal of the mean (i.e., normally 0.05).

Pedigo (1989) suggests the formula n=(ts/Ei)2, where the t value is found usually at 0.05 probability. Wilson (1985) presents a generalized form which can be used for organisms having distribution patterns ranging from highly clumped to uniform: n=t /2D2.a.ib-2; where t2/2 is the standard normal variate for a two tailed confidence interval, D is a fixed proportion of the mean and is used to define 1/2 of the confidence interval (1/2 C.I. = DX), and a and b are Taylor's coefficients.

Seasonal distribution

Biological environments are characterized by having annual cycles of resources and unfavorable conditions. Ecologically, insects have developed a phenological "strategy" to adapt their life cycles to these environmental cycles. An understanding of this seasonality relationship is required to develop efficient pest management programs.

A knowledge of the seasonal distribution of the dominant pest and beneficial species in the agroecosystem will determine, if necessary, when to apply appropriate control








48

tactics. The seasonal distribution of some insect species varies according to several factors such as weather, geographical areas, crop, season, etc. However, other insect species follow typical seasonal distribution patterns among years, some examples are the pea aphid Acvrthosihon Pisum (Maiteki et al. 1986), Nabis spp. (Shepard et al. 1974) and Geocoris spp. (Funderburk and Mack 1987) in soybean. Unfortunately, not all the species follow a typical predictable distribution pattern (Carner et al. 1974).

Harding (1961), after studying thrips infestations in onions in South Texas, concluded that: a) no concentrated movement or thrips into onion fields occurred, b) thrips movement into onion fields lessened as other host plants became abundant and the onions approached maturity, c) precipitation and mean daily temperatures below 50 degrees F. reduced thrips movement, and d) destructive infestations resulted from build up by breeding in the field and not from thrips movement from outside of the field into the field.

Race (1965) found a direct relationship between populations of E. occidentalis before and after cotton plantings. With this data, post-planting populations could be predicted with accuracy. By observing the seasonal abundance records of E. occidentalis, He concluded that there was not a typical pattern for the seasonal distribution of the populations during the five years of his study.








49

Watts (1936) found that both adult and larval populations of E. tritici were small in early and mid-winter (December 20 to March 20) in South Carolina. With the beginning of spring (March 20 to June 20) the population ordinarily remained small and thrips activity minimal. High population densities were reached in the summer (June 20 to September 20) followed by a final gradual decline.

The arrangement of combined.crops in the field can also affect thrips populations. Letourneau and Altieri (1983) found that the population of E. occidentali was much greater on squash leaves in monoculture than in tricultures with corn and cowpea intercropped with squash. The minute pirate bug, Orious tristicolor, a thrips predator, exhibited a more rapid colonization rate in tricultures than in monocultures. Binomial (or presence/absence) samplin Dproram

The direct counting of abundant individual organisms, many of which are too small to be detected without the aid of a microscope, is usually a slow, arduous and costly process. Indirect estimates of insect population densities, which eliminate the necessity for counting every organism per sample and relies solely upon the presence or absence of individuals in sampling units, has been developed (Gerrard and Chiang 1970, Wilson and Gerrard 1971, Gerrard and Cook 1972). The basic idea is quite simple: if the frequency distribution of an organism can be identified, then mean density can be








50

estimated from the proportion of samples attaining or exceeding some specified density.

If the dispersion of the population in a particular habitat can be described by the negative binomial distribution with a known k, the probability of a particular mean population can be estimated by the formula: P=k[(/l1-P')/k-l]; where PY is the mean population, P' the probability of the insect being present in the sampling unit determined on the basis of presence or absence sampling, and k the dispersion parameter, a measure of the amount of clumping in negative binomial distribution: K=(. ) /(s2-X). It is reliable only if the critical density levels are related to values of p' less than about 0.8; above this uncertainty associated with the predictions becomes too great (Wilson and Gerrard 1971, Southwood 1978).

Wilson and Room (1983) incorporated Taylor's power law into a mathematical function relating the proportion of sample units having organisms present to the variance and mean per unit. Based on this relationship the control status or density of the species can be assessed.

Binomial sampling programs have been developed for small and usually highly aggregated organisms such as mites (Wilson et al. 1981, Zalom et al. 1984, Nyrop et al. 1989), aphids (Wilson et al. 1983), larvae of the European pine sawfly (Wilson and Gerrard 1971), and corn root-worm egg populations (Gerrard and Chiang 1970).








51

Factors affecting estimation of insect densities

Several variables that can affect the estimation of insect population are usually not considered as required components in sampling programs. These factors can cause disparity among observations of various researchers and can result in serious mistakes being made when sampling data is being taken for insect pest management purposes. Time of day of sampling, hedge effect, neighboring vegetation, wind speed and direction, height and/or position of the sample unit in the plant, and other factors, are variables that can easily induce variability in the results of sampling and should be also considered in the development of sampling programs.

Hedge row and neighboring fields can affect the diversity of the fauna and the distribution of species in the crop field. Hedge row may have an especially rich flora of woodland shrubs and trees with various grasses and herbs or other crops which allow them to support more species of insects than neighbouring crops. Many species move between these two habitats especially from hedges to crops. Lewis (1969a), studying insect communities on vegetation in a mixed hedge row and in neighbouring fields of pasture and field beans, found that diversity of species was always greater in the hedge row, but fluctuated more than in the beans and short pasture, as different flowers bloomed.

The distribution of individuals for specific species within the field has also been shown to be affected by hedge








52

row. Lewis (1969b) found that most'species populations were concentrated near the hedge row. Species of some families such as Empididae, Miridae, Syrphidae and some Lepidoptera were concentrated on the hedge row. Limothris sp. and other species were also affected by this factor in minor but significant extent. Both wind and neighboring flora were responsible for this distribution. Similar results are reported for the tomato pinworm, Keiferia lvcopersicella (Walsingham) by Pena (1983), and for Liriomvza trifolii in tomato (Price et al. 1981).

Sampling time is another variable that can affect insect populations. This variable can be induced for several factors such as temperature, humidity, light, wind, and others. Tappan (1986a), studying the relationship of sampling time to E. fusca numbers in peanut foliage buds and flowers, found no effect on number per bud, but the number per flower increased significantly between 0800 and 1000 hours with the largest number occurring between 1100 and 1200 hours. Watts (1936) found that the greatest activity of E. tritici in cotton seedlings occurred between 1000 and 1630 hours. He also found that activity declined under higher humidity and lower temperatures near nightfall. Under similar conditions the female, which is larger, was decidedly less active than the smaller and more slender male.








53

Economic Iniurv Level and Economic Thresholds

The economic injury level (EIL) was initially defined by Stern et al. (1959) as the level at which damage can no longer be tolerated and therefore the level at or before which it is desirable to initiate deliberate control activities. This concept was developed largely as a means for more rational use of insecticides. The same authors defined the economic threshold (ET) as the population density at which control measures should be applied to prevent an increasing pest population from reaching the EIL. Both concepts have been criticized by several authors and even other names have been proposed and/or used. Pedigo et al. (1986), for example, considers that if the economic injury level is defined as a population density rather than an injury level, the EIL should be actually called "critical population density". However, both concepts are widely accepted and remain as some of the most pervasive and influential elements in agricultural pest management.

According to Pedigo et al. (1986), four primary components affect the EIL: market value, management cost, injury per insect unit density, and host damage per unit of injury. The mathematical relationship of these components have been widely defined.

Economic damage, according to Stern et al. (1959), is expressed as a monetary value and occurs when:








54

C(a)
Pedigo et al. (1986) suggested a different model for use in practical insect management: EIL=C/VID; where ZIL is number of injury equivalents per production unit (e.g. insects/ha, all of which live to attain their full injury potential), C cost of the management activity per unit of production (e.g.

$/kg), I injury units per insect per production unit (e.g. proportion defoliated/ (insect/ha)) D damage per unit injury (e.g. kg reduction/ha/proportion defoliated)

In spite of the many advantages of the EIL, it also has

several limitations related to the type of pests or injury that can be addressed, the control tactics used, the research requirements, and the existence of many pest species and variable environments.

As a response to the relative unsuitability of the EIL for multiple pests, Hutchins et al. (1988) proposed a new technique for developing and using multiple-species economicinjury levels. The theoretical basis for this approach consists of grouping insects into injury guilds based on the plant's response to the injury. This requires that injuries by the different species of interest produce a homogeneous response in the plant.








55

Economic thresholds for flower thrips vary greatly between crops and depending upon the presence of TSWV. Few thresholds have been experimentally determined. Others have been adopted in extension programs based on experience. A threshold of more than 100 thrips per bloom has been adopted for E. occidentalis in cotton blooms in several states to prevent poor pollination, blossom drop and small boll shed (Sprenkel 1987, Graves et al. 1987). However, a much lower threshold (2 thrips per plant) is used for the same species in the same crop but at seedling time (Race 1961). Carlson (1966) experimentally determined an economic threshold of 10

- 20 adult E. occidentalis per bud on safflower plants. An economic threshold of 10 thrips per flower or florete (900 to 1000 flowers per head) was determined by Carlson (1964a) in onion seed plants. However, populations up to 3.6 thrips per flower increased pollination. Yokoyama (1977) found that clusters that supported up to 1582 adult and nymphs of E. occidentalis did not have a greater amount of surface scars than non infested clusters of seedless table grapes. "Detectable" levels of E. occidentalis constitute an economic threshold in ornamentals if TSWV is present but, if the disease is absent, thrips can be tolerated up to a level where premature aging of blossoms occur, depending on the plant species (Sprenkel 1987). Populations of about 10 E. bispinosa thrips per flower increased bloom drop on tomato (Schuster and Price 1987). An average of 1 to 2 thrips per flower has been








56

determined through experience to be the threshold for Z. occidentalis in tomato for fresh market in the south east (Sprenkel 1987). However, these numbers could be much lower if TSWV is present.












CHAPTER 3

SEASONAL PATTERNS OF Frankliniella spp.
(THYSANOPTERA:THRIPIDAE) IN TOMATO FLOWERS AND
INFLUENCES OF SEVERAL FACTORS ON SAMPLE ESTIMATES Introduction
Flower thrips, Frankliniella spp., are polyphagous insects that are pests of numerous crops worldwide (Cho et al. 1989). The western flower thrips, E. occidentalis (Pergande), and the tobacco thrips, E. fusca (Hinds), are pests of tomato because they vector tomato spotted wilt virus (Sakimura 1962, 1963). Adult female E. occidentalis oviposit in small tomato fruit, thereby causing cosmetic damage (unpublished data).

Although originally confined to the western half of the United States, E. occidentalis has moved into the Southeast, being first recorded in Georgia in 1981 (Beshear 1983). This species is now present throughout most of the region, including Florida (unpublished data). Tomato spotted wilt virus has a wide host range and is worldwide in distribution (Cho et al. 1989), but has only recently been found in the Southeast. The disease was reported extensively in the region during 1986 on tomato, pepper, and other crops (Reddick et al. 1987, Hagan et al. 1987). It was first noted in Florida in 1986 (Olson and Funderburk 1986).



57








58

Efforts to develop strategies to manage thrips in tomato in the Southeast are hampered by a lack of important biological information. The species composition and seasonal

abundances of thrips inhabiting tomato in the region are unknown. The primary purpose of this study was to determine seasonal abundance of thrips inhabiting tomato flowers. Multiple cropping seasons are typical of tomato production in the Southeast, with tomatoes in North Florida grown during the spring and fall. Consequently, seasonal abundance of thrips was determined for both cropping seasons. Another objective was to evaluate the effects on density estimates of sample location within a field, sample location on individual tomato plants, and the time of day when sampling. This information is needed to develop sampling programs to estimate density for scouting purposes.

Materials and Methods
Thrips densities in tomato flowers were estimated weekly in tomato fields during the spring growing seasons of 1987, 1988, and 1989 and during the fall growing seasons of 1987 and 1988. Two fields about 5 ha in size located in Gadsden County, Florida, were sampled. Each was a production field that was managed in a manner typical for the agroecosystem. Sampling in each field in all years began within one week of first blooming and continued until near final harvest.

Flower samples for estimating thrips densities were placed in vials containing 70% ethyl alcohol and returned to








59

the laboratory for further processing to determine the number of adults of each thrips species and the total number of adult and immature thrips. The contents of each vial was transfered to a 4-cm-diam petri dish and examined using 6.5 to 40X magnification. A microscope slide was prepared for each adult thrips, with CMC-10 (Masters Chemical Co., Inc., 520 Bonnie Lane, Elk Grove Village, Illinois 60007) used as the clearing medium. After at least 24 hr, adults were identified using 100 to 1000X magnification. Adult thrips were identified, by using a key to the genus Frankliniella developed by R. Beshear (unpublished data) (Table 2.1 and Fig. 2.1).

During the spring and fall cropping seasons of 1987, the sampling protocol was designed with the sole objective of determining seasonal abundance. A total of 20 random samples of individual flowers were collected weekly in each field. During the spring and fall cropping seasons of 1988 and the spring cropping season of 1989, samples were collected to determine seasonal abundance and to evaluate the effects of several factors on density estimates. A total of 32 samples of individual flowers were collected weekly in each field, with the samples taken so that analysis of variance (ANOVA) procedures were used to evaluate the effects on density estimates of sample location within a field, sample location on individual plants, and the time of day when sampling. Consequently, the 32 samples taken in a field on each sample date were subdivided so that 4 random samples were taken from








60

each of 8 treatments. Treatments were a factorial arrangement of two locations within the field (i.e., plants located < 40m from the field edges and plants > 40m from the field edges), two locations on the plant (i.e., upper and lower half of each sampled tomato plant), and two times during the day when taking samples (i.e., 0800 to 1000 hr and 1100 to 1300 hr). Data from the spring growing seasons of 1988 and 1989 were used to evaluate these effects on density estimates of adults of each species and the total number of adults and immatures.

Data for each growing season were analyzed as a splitplot randomized complete block over time, with treatments as whole plots and dates as subplots (Steel and Torrie 1960). Individual fields were treated as block effects. The treatment sum of squares in each analysis was further divided into the main and interactive effects of sample location within fields, sample location on individual plants, and time of day when sampling.

The mean number of adults of each species and of total adults and immatures per flower was determined for each sample date by averaging over all individual samples. The mean number of flowers per plant on each sample date also was estimated in each field during the 1988 spring and fall growing seasons and the 1989 spring growing season. For these estimates, the number of flowers on 6 randomly selected tomato plants was determined on each weekly sample date. The average number of thrips per plant was obtained then by multiplying the mean








61

number of thrips per flower by the average number of flowers per plant.

Results and Discussion

Adult thrips accounted for about 88% of the total thrips (n=2353) collected in the tomato flowers during the three years of this study. About 97% of these adults were E. occidentalis, F. tritici, and E. fusca. Most of the remaining adults were E. bispinosa, with other species accounting for less than 1% of the total collected. These other species of adult thrips included Thrips tabaci Lindeman, Pseudothrips ineualis (Beach), Plesiothrips perplexus Beach, Chirothris sp., and Sericothrips sp.

Adult thrips were present during each spring cropping season (Fig. 3.1). Greatest densities occurred during May each year. Species included E. occidentalis, E. tritici, and E. fusca. Populations of E. occidentalis and E. tritici were abundant each year. Populations of E. fusca were present during two of the spring cropping seasons.

Adult thrips were not present on most samples dates during the fall cropping seasons (Fig. 3.2). The species occurring in the fall included E. tritici and E. occidentalis. Populations of E. tritici occurred each year. Populations of E. occidentalis were present only during 1987.

Adults of E. occidentalis, E. tritici, and E. fusca rarely inhabited other tomato plant structures (unpublished data). These species apparently prefer only the tomato flower,








62

although insecticides used to control lepidopterous and other pests on tomatoes may have reduced any populations inhabiting other plant structures. Estimates of the number of flowers per plant on each sample date during the spring and fall growing seasons of 1988 and the spring growing season of 1989 were used to convert the number of adult E. occidentalis, E. tritici, and E. fusca per flower to the number of thrips per plant (Fig. 3.3). For each species, seasonal patterns of abundance per flower were significantly (2 < 0.001) correlated to seasonal patterns of abundance per plant. Correlation coefficients during the spring of 1988 for E. occidentalis, E. tritici, and E. fusca were 0.97, 0.99, and 0.97, respectively. The correlation coefficient during the fall of 1988 for F. tritici was 0.99. Correlation coefficients during the spring of 1989 for E. occidentalis, E. tritici, and E. fusca were 0.96, 0.98, and 0.98, respectively.

The seasonal abundance data shows that thrips populations are more abundant at certain times of the year, with population trends apparently unrelated to crop phenology or the number of flowers per tomato plant. Populations of E. occidentalis, F. tritici, and E. fusca were common between late April and early June, with greatest densities during May. Densities were very low on the other sample dates, especially during each fall cropping season. For each species, seasonal patterns of density per flower were very similar to seasonal patterns of abundance per plant.








63

The effects of field position, plant position, and time of sampling on density estimates were determined using data from the spring growing seasons of 1988 and 1989 (Table 3.1). Seasonal densities of all species were low during the fall growing season of 1989, thereby eliminating the possibility of examining the effects of the above treatments on density estimates during that growing season. Plant position of sampling had a statistically significant effect on density estimates of adult E. occidentalis, adult E. tritici, total adults, and immatures (Fig. 3.5). Densities of these adult species were greater in flowers located on the upper half of the plants than in flowers on the lower half of the plants.

Adult thrips, therefore, prefer the flowers on the upper half of the plants. Conversely, density estimates of immature thrips were greater for flowers located on the lower rather than upper half of the plants. Because of the very low numbers of immature thrips in all flowers, it is believed that insecticides sprayed in the fields greatly reduced their populations. There may have been better immature survival on the lower flowers, because of poorer coverage of insecticides.

Sample location within a field significantly affected density estimates of E. occidentalis during 1989, with greater populations near field margins (Table 3.1, Fig 3.4). This difference was not significant in 1988, but a singnificant field position * plant position and field position * time of sampling interactions revealed that field position affected








64

density estimates. According to these interactions, density estimates on the upper half of tomato plants were greater on field margins than in the center of fields. Density estimates in marginal areas were greater between 1100-1300 hours than at 0800-1000 hours. Other adult species or inmmatures were not affected by location of samples within a field (Fig. 3.4). Time of sampling had no significant effect on density estimates of adults of any species, total adults, or total immatures (Fig. 3.6).

The results from these studies provide important information for integrated pest management programs. Economically damaging populations of thrips species capable of transmiting TSWV or causing cosmetic fruit damage occur most frequently during the spring cropping season. In North Florida, this period occurs between late April and early June. It should be possible to focus detection and management efforts to this period of the spring growing season.

This data indicates that estimates of thrips densities in the flower also can be used to estimate thrips densities per plant. The three common species of thrips in tomato apparently are primarily flower inhabiting. However, numerous pests reach outbreak densities frequently in tomatoes, and insecticides are required as an obligate control measure. Insecticides sprayed in the commercial fields sampled in this study may have controlled any populations of thrips inhabiting other plant structures but not adult thrips in the flowers.








65

The results involving the effects of several factors on density estimates of thrips provide important information when monitoring thrips in tomatoes. The time of day when sampling had no significant influence on density estimates of any species. Field position affected densities of . occidentalis only. The location of flowers on plants greatly affected densities of F. occidentalis and E. tritici, but not Z. fusca. These effects on density estimates of some species will need to be considered in scouting programs. Separate estimates of density of E. occidentalis will need to be made for marginal and non-marginal areas of tomato fields. Also separate estimates of density of f. occidentalis and E. tritici will be needed for flowers located on the upper or lower half of tomato plants.







66




3
-F. occidntalis 1987
2 ----- F. tritici
- F. fusca


0

3

1988
2 1



3

1989
2


1 /


April May June July


Fig. 3.1. Mean number of adult E. occidentalis, E. tritici,
and F. fusca per flower in tomato fields sampled weekly during the spring of 1987, 1988, and 1989
in Gadsden County, Florida.








67









3
.----- F tritil 1987
2 - F. occientails








3

1988
2






0 I
Aug Sept Oct Nov








Fig. 3.2. Mean number of adult E. occidentalis and E. tritici
per flower in tomato fields sampled weekly during
the fall of 1987 and 1988 in Gadsden County,
Florida.









68






160
-FL- u FALL 1988 F. odoUWtaI 100 -- F. No


50


AUG SEPT OCT ' NOV


150

SPRING 1988
100



60



APRIL MAY JUNE JULY


150

SPRING 1989

100



60


0
APRIL MAY JUNE JULY

Fig. 3.3. Mean number of flowers and adult thrips per plant
in the tomato fields sampled weekly during the
springs of 1988 and 1989 and the fall of 1988 in
Gadsden County, Florida.













Margin M Center Margin Canter



1988 1989


1. 1








0.6 - 0.5





0 0 F.occld. F.fusca F.tritici Adults Immatures F.occid. F.fusca F.trltici Adults Immaturee

Fig. 3.4. Effect of field position of sampling on density estimates of Frankliniella
spp. thrips in flowers in two tomato fields sampled during the spring in
Gadsden County, Florida. (** indicates significantly different at the
probability level of 0.01 according to an analysis of variance).












2 2
Upper half M Lower half Upper half M Lower half



1988 1989






1 "1 0

**


**
0.5 0.5





0 0
F.occid. F.fusca F.trltici Adults Immaturee F.occid. F.fusca F.trltlci Adults Immatures
Fig. 3.5. Effect of plant position of sampling on density estimates of Frankliniella
spp. thrips in flowers in two tomato fields sampled during the spring in
Gadsden County, Florida. (* and ** indicates significantly different at the
0.05 and 0.01 probability, respectively, according to an analysis of
variance).













2 2
8 - 10 hrs M 11 - 18 hr M 8 - 10 hrs M11 - 13 hra



1988 1989






1.6 1.6





0.5 - 0.50 0
F.occid. F.fusca F.trltlci Adults Immatures F.occld. F.fusca F.triticl Adults Immatures
Fig. 3.6. Effect of time of sampling on density estimates of Frankliniella spp. thrips
in flowers in two tomato fields sampled during the spring in Gadsden County,
Florida. (No means are significantly different at the 0.05 level of
probability according to an anlysis of variance).






Table 3.1. Effects of field position, plant position, and time of sampling on seasonal
densities of adult F. occidentalis, E. fusca, and E. tritici and total adult and immature thrips in flowers in two tomato fields sampled during the spring
1988 and 1989 growing seasons in Gadsden County, Florida.



Treatments Mean number of thrips per bloom
Field Plant Time of F. occidentalis F. fusca F. tritici Adults Immatures Position Position Sampling - 18 989 1988-T99 1-88 1989 1988 1989 1988 1989


Upper 8:00-10:00 0.69 1.04 0.24 0.06 0.42 0.95 1.36 2.05 0.04 0.17
Margin half 11:00-13:00 1.11 1.45 0.07 0.04 0.52 0.63 1.69 2.11 0.10 0.08
(<40m from
margin) Lower 8:00-10:00 0.34 0.50 0.07 0.03 0.06 0.08 0.46 0.61 0.21 0.32
half 11:00-13:00 0.57 0.71 0.07 0.06 0.08 0.06 0.71 0.83 0.20 0.25

Upper 8:00-10:00 0.71 0.63 0.18 0.05 0.50 0.94 1.39 1.61 0.10 0.10
Center half 11:00-13:00 0.47 0.72 0.13 0.02 0.55 0.86 1.14 1.60 0.10 0.10
(>40m from
margin) Lower 8:00-10:00 0.58 0.32 0.16 0.05 0.07 0.08 0.81 0.46 0.28 0.20
half 11:00-13:00 0.57 0.39 0.13 0.01 0.03 0.06 0.72 0.46 0.08 0.32

Analysis of variance df F value

Field position (F) 1 2.10 14.19** 0.81 2.76 0.04 0.38 0.12 5.65 0.00 0.43 Plant position (P) 1 13.01**19.24** 1.26 0.31 29.19**69.06** 41.98**65.15** 10.21* 19.54** Time of sampling (T) 1 2.32 3.11 2.47 2.76 0.15 1.39 0.32 0.19 1.24 0.05 F x P 1 10.96* 2.18 0.81 0.00 0.20 0.38 3.78 0.45 0.69 0.01 F X T 1 11.96* 1.10 0.20 4.88 0.11 0.45 4.30 0.22 3.26 3.83 P x T 1 0.03 0.28 1.26 1.24 0.25 0.91 0.03 0.07 3.78 0.89 P x Tx F 1 2.56 0.15 0.81 2.76 0.00 0.45 0.32 0.05 0.95 0.64

* Means statistically different at 0.05.

** Means statistically different at 0.01.













CHAPTER 4

BINOMIAL SAMPLING PROGRAM
FOR Frankliniella spp. IN TOMATO
ACCORDING TO THE POSITION OF SAMPLING IN THE PLANT Introduction

Several species of flower thrips, Frankliniella spp., are known to inhabit tomato plants and have been associated with cosmetic fruit damage (unpublished data) and/or transmission of tomato spotted wilt virus (TSWV) (Sakimura 1962, 1963). Direct damage to tomato fruit was first recognized in North Florida as an economic problem in 1985 about the time E. occidentalis was first recorded in the tomato crop (Olson and Funderburk 1986, Funderburk 1988). E. tritici, E. fusca, and E. bispinosa have been also found to inhabit tomato flowers (Salguero and Funderburk 1989). The cosmetic damage produced by flower thrips on tomato fruit has resulted in rejections of fruit or lowering of grade by regulatory authorities (personal observations).

TSWV is a serious disease affecting the production of tomato and other food and ornamental crops worldwide (Cho et al. 1989). Economic losses in tomato due to TSWV have been reported in Canada (Allen and Broadbent 1986), Hawaii (Cho et al. 1986), and Louisiana (Greenough et al. 1985). This disease was first detected and confirmed in tomato in North Florida


73








74

in 1986 (Sprenkel 1988). The presence of Frankliniella spp. and TSWV in this region has resulted in an increase in both dosages and frequency of insecticide applications.

The focus of integrated pest management programs is to

estimate population densities of individual species and apply control tactics only if economic thresholds are reached. However, the direct counting of abundant and small individual organisms is usually a slow, arduous, and costly process. Binomial sampling is a practical and reliable procedure which eliminates the necessity for counting every organism per sample unit and relies solely upon the presence or absence of individuals in each unit. This technique is based upon the relationship between the proportion of sample units with one or more organisms (incidence) and the density of insects (mean) per sample unit. A sample estimate of incidence is then used to predict density (Gerrard and Chiang 1970, Wilson and Gerrard 1971, Gerrard and Cook 1972).

Sampling programs to practically and reliably estimate densities of flower thrips in tomato have not been developed and are needed to effectively manage these pests. Direct counting of flower thrips is technically difficult and a slow and arduous process, because of their small size and their tendency to quickly escape when disturbed. However, Frankliniella spp. thrips presence or absence in flowers apparently can be determined directly in the field. Consequently, binomial sampling may be a feasible field-








75

sampling method for estimating Frankliniella spp. densities in tomato fields.

The purpose of this study was to develop a sampling technique suitable for scouting programs to estimate density of flower thrips in tomato flowers. Dispersion characteristics of Frankliniella spp. were quantified and the data used to develop a binomial sampling program.

Materials and Methods

Population densities of flower thrips were estimated weekly in two tomato fields during the spring growing seasons of 1988 and 1989 and during the fall growing season of 1988. Each field was about 5 ha. in size and located in Gadsden County, Florida. Each field was managed in a manner typical for tomato production fields. Sampling began when first flowers became available and continued until near final harvest. A total of 13 and 12 samples were taken in each field during the spring and fall growing season, respectively. Each sample consisted of 32 sample units (individual open flowers). Flowers were placed in individual vials containing 70% ethyl alcohol and returned to the laboratory for processing. The content of each vial was transferred to a 4-cm-diam petri dish and examined using 6.5 to 40X magnification. A microscope slide was prepared for each adult thrips, with CMC-10 (Master Chemical Co., Inc., 520 Bonnie Lane, Elk Grove Village, Illinois 60007) used as the clearing and preserving medium. Adults were identified 24 hours later using 100 to 1000X








76

magnification. Adults were identified by using a key to the genus Frankliniella (Table 2.1, Fig. 2.1) developed by R. Beshear (unpublished data)'. The number of adults of each thrips species and the total number of adults and total number of immatures in each sample were determined.

The 32 sample units (individual blooms), taken in a field on each sample date, were subdivided into 8 treatments with 4 replicates for each treatment. Treatments were a factorial arrangement of the following three factors: two locations of sampling within the field (i.e.; margin, plants located < 40m from the field edges; and center, plants located > 40m from the field edges), two position of sampling on the plant (i.e.; upper and lower half of each tomato plant), and two times during the day when taking samples (i.e.; 0800 to 1000 and 1100 to 1300 hr).

Plant position of sampling was the only factor which significantly affected density estimates of F. occidentalis and E. tritici during each year (refer to Chapter 3). No factor significantly affected density of F. fusca. Taylor's power law relationships (Taylor et al. 1978) for each species therefore were determined for the upper half and the lower half of tomato plants and for data pooled over both plant positions. The same relationships also were used to describe dispersion characteristics of total adults and total immatures.








77

The Taylor's power law relationships were used to develop the binomial sampling program for adult E. occidentalis, E. tritici, and F. fusca inhabiting the upper half of tomato plants. The slope and intercept of the Taylor's power law relationships were used in the Wilson and Room (1983) equation to determine the relationships between mean density and the proportion of infested tomato flowers. The number of samples needed to estimate thrips densities at the 10 and 25% precision levels was determined by using the equation in Ruesink (1980).

Results and Discussion
Most of the thrips collected in the tomato flowers through both growing seasons were adults, with only 12% being immatures. Three species accounted for about 97% of the adults collected, including E. occidentalis, E. tritici, and E. fusca (Refer to Chapter 3). Plant position of sampling significantly affected density estimates of adult E. occidentalis and E. tritici, with densities much greater on flowers located in the upper half of the tomato plants than on flowers located on the lower half of the tomato plants. Consequently, regression statistics of Taylor's power law relationships for individual sample estimates of E. occidentalis, E. tritici, and E. fusca adults and of total adults and total immatures according to plant position of sampling are given in Table 4.1. Taylor's power law relates variance (s2) to mean density (m) by the relationship s2=amb. Taylor et al. (1978) considered the slope








78

(k) to be a constant for a species (with values of ] < 1, b = 1, and 2 > 1 indicating uniform, random, and aggregated distributions, respectively) and the intercept (A) to be reflected by sample unit size. Taylor's power law allows for a description of a species distribution pattern in relation to density.

For nearly all of the thrips regression relationships, b was statistically > 1 (P <0.001) for a t test. These values of ] indicate that E. occidentalis, E. tritici, and total adults and immatures thrips were aggregated over a wide range of densities in tomato flowers. E. fusca showed an aggregated dispersion in 1988 (2 = 1.66), but its dispersion in 1989 was random (b = 0.98). Apparently, E. fusca dispersion characteristics were slightly aggregated to random.

Determination of the functional relationship between the proportion of infested tomato flowers and thrips density allows for estimation of thrips densities by presence or absence sampling. The functional relationships between the proportion of infested flowers and density were determined for E. occidentalis, E. tritici, and E. fusca adults inhabiting flowers located on the upper half of the tomato plants. Estimating thrips densities in flowers on the upper half of the tomato plants is expected to be important for management purposes.

For each flower thrips species, the relationship between the proportion of infested tomato flowers and density were








79

similar at densities of about 1 thrips per flower (Figure 4.1). For example, a proportion of infested tomato flowers of 0.65, 0.66, and 0.67 represented a mean density of 1.1 adult thrips per flower for �. tritici, E. occidentalis, and E. fusca, respectively. Populations of . occidentalis and E. tritici in tomato flowers were very aggregated and densities of these species greater than 1.4 thrips per flower could not be reliably estimated by the proportion of infested tomato flowers. Because populations of E. fusca are less aggregated than populations of E. occidentalis and E. tritici, the binomial sampling program could be used to estimate densities up to about 2 thrips per flower.

Because adults of each species did not exhibit the same dispersion patterns, combining counts of all species in the flowers to estimate total density of thrips would not be statistically reliable. .Further, each species differs in economic importance. E. occidentalis and E. fusca transmit TSWV, but E. tritici does not (Sakimura 1962, 1963). Cosmetic fruit damage is caused by oviposition by E. occidentalis, with E. tritici and E. fusca apparently not causing cosmetic damage (refer to Chapter 5). For these reasons, estimates of total thrips densities in flowers has limited management value in situations where each species are present.

The number of samples needed to estimate density of E. occidentalis and Z. fusca. in tomato fields by using the binomial sampling program was determined for the 10 and 25%








80

levels of precision (Figure 4.2). The number of samples needed to estimate density of F. tritici could not be calculated using the Ruesink (1980) equation because B values were greater than 2. The number of samples needed to obtain a 10% level of precision was above 100 for densities lower than 1 thrips per flower for E. occidentalis and E. fusca. About 90 samples were needed for densities above 1 thrips per flower for F. occidentalis. Fewer samples were needed at high densities for E. fusca, because its populations are less aggregated than E. occidentalis.

The number of samples needed to estimate densities of E. occidentalis and E. fusca at the 25% level of precision was between 15 and 20 for densities above 1 thrips per flower. The number of samples needed to estimate densities of E. occidentalis at densities higher than 1 thrips per flower was about 15. The number of samples needed to estimate densities of E. fusca was lower.

A 25% level of precision usually is considered acceptable for extensive sampling programs. To obtain this level of precision, fewer than 20 presence or absence samples are needed to estimate densities of E. occidentalis or E. fusca. Consequently, the binomial sampling program appears feasible to estimate individual densities of adult thrips species. However, the implementation of this sampling program in areas where two or more species of Frankliniella occur simultaneously probably will not be practical. Because adults








81

of each species did not exhibit the same dispersion patterns, combining counts of all species and quantifying dispersion characteristics of all species is not statistically valid. Estimates of density of all species combined has limited management value anyway, because each species differs in economic importance. E. occidentalis and E. fusca transmit TSWV but E. tritici does not (Sakiura 1962, 1963). Cosmetic fruit damage occurs when E. occidentalis oviposits on small tomato fruit, with E. tritici and E. fusca apparently not responsible for cosmetic fruit damage (refer to Chapter 5). Also, it is not possible to reliably identify Frankliniella spp. thrips in the field. Because several species of thrips commonly inhabit tomato flowers in North Florida, the use of a binomial sampling program will have little value for integrated pest management programs. However, the binomial sampling program will be useful for scouting programs in geographical areas where only one species commonly inhabit tomato flowers.






Table 4.1. Regression statistics of Taylor's power law relationships for
Frankliniella spp. thrips sample data in tomato fields in North
Florida during 1988 and 1989.


Plant position of samDlin

UDDer half Lower half UDDer & lower half


Species Year b a r2 b a r2 b a


E. occidentalis 1988 2.26 0.66 0.96 2.02 0.78 0.95 2.22 0.72 0.97
1989 1.80 0.91 0.99 1.77 0.89 0.97 1.92 0.85 0.99 . fusca 1988 1.70 0.93 0.98 1.24 0.98 0.94 1.66 0.93 0.96 1989 1.05 1.00 0.99 0.90 1.00 0.99 0.98 1.00 0.99 E. tritici 1988 2.33 0.83 0.91 1.32 0.98 0.91 2.88 0.78 0.90
1989 2.14 0.95 0.98 0.98 1.02 0.89 2.59 0.91 0.99 Adults 1988 2.25 0.62 0.93 1.91 0.81 0.92 2.26 0.71 0.92
1989 1.76 0.87 0.96 1.82 0.78 0.98 1.98 0.95 0.96 Immatures 1988 2.07 0.89 0.89 3.73 0.87 0.71 3.03 0.95 0.70
1989 1.87 0.93 0.83 2.78 0.85 0.85 2.87 0.91 0.84


All relationships are significantly linear beyond the 0.001 level. b = slope, a = intercept.








83







1 - ---0 .9 - ..... ...,


P 0.8
r


0
r
i , i -- --!:---------- i !i

n 0.5 -i ,! i


n 0.4f
e
s 0.3t
e
.0.. .. .............. .. .. . ....... .......
d 0.2 F. occidentalis

.0.1- - F. tritici.
---F. fusca


.5 1.1 2.0 3.0 4.1 No. thrips per flower
Fig. 4.1. The relationships between the proportion of infested flowers and Frankliniella spp. density per tomato
flower as estimated by using the Wilson and Room
(1983) equation.










84











200

176 jPrecisio e.

160 .........



S76 , i
100













- F. ooodentalis 30 ..i.l... F. .uso.
30




2 0 ............







1 0 0 ...... . ...



1 2 3 4 6 6

No. thrips per flower

Fig. 4.2. The relationships between the number of samples
needed to estimate density at the 10 and 25%
precision levels and the number of thrips per
flower for Frankliniella spp. in tomato as
determined by using the Ruesink (1980) equation











CHAPTER 5

FLOWER THRIPS, Frankliniella spp.
DAMAGE TO TOMATO FRUIT

Introduction
Flower thrips, Frankliniella spp., are important pests in agriculture because of the capacity of some species for transmiting tomato spotted wilt virus (TSWV) and the direct damage when feeding or ovipositing on foliage, flowers, or fruits. Because of the wide host range of Frankliniella spp. (and TSWV), serious economic losses occur on cotton, peanuts, onions, and other crops in the southeastern U. S. (Sprenkel 1987) and on tomato in Canada and Hawaii (Allen and Broadbent 1986, Cho et al. 1989). However, flower thrips have not been considered important pests of tomato in the continental U. S. (Lange and Bronson 1981, Pohronezny et al. 1989).

TSWV was only recently found in the southeastern U. S. on tomato, peppers, and other crops (Kucharek 1986, Olson and Funderburk 1986, Hagan et al. 1987, Reddick et al. 1987, Sprenkel 1988). Simultaneously, flower thrips were implicated as the cause for cosmetic fruit damage occurring on tomatoes grown in North Florida (Olson and Funderburk 1986, Funderburk 1988). This cosmetic damage consisted of small dark scars or indentations usually surrounded by a lightened or whitish skin area. A similar damage, caused by oviposition of the western


85








86

flower thrips, E. occidentalis, was reported on grapes in California (Jensen 1973, Stafford 1974, Yokoyama 1977) and apples in Arizona (Terry and DeGrandi-Hoffman 1988). Although tomatoes are California's most important vegetable crop (Anonymous 1985) and E. occidentalis is present and common, cosmetic damage to the fruit has not been reported. The same condition applies for Canada and Hawaii.

The cosmetic fruit damage on tomato has become a serious economic problem in North Florida. Rejection or lowering of grade of damaged fruit has occurred and insecticide applications increased. An understanding of the cause of the damage on the tomato fruit and its association with Frankliniella spp. is necessary to implement management programs. The major purpose of this study was to determine if the Frankliniella spp. thrips inhabiting tomato flowers cause cosmetic damage to tomato fruit. The relationship between the number of scars per fruit and the thrips densities on tomato flowers and small fruit also was quantified.

Materials and Methods
Several densities of adult female flower thrips (0, 1, 2, 4, 8, and 16) of three species (E. occidentalis, E. fusca, and E. tritici) were confined on individual small tomato fruit and flowers. Tomato plants for this experiment were grown in a greenhouse at about 270C at the North Florida Research and Education Center, Quincy, Florida, from March to July 1989. Small fruits and open flowers were infested with E.








87
occidentalis and E. fusca, while only open flowers were infested with E. tritici. More than 500 experimental units were established but about 25% of the infested flowers aborted (affecting all treatments). This abortion occurred mainly during the last days of the study when temperatures were often greater than 270C. Insecticides to control thrips were applied when natural infestations of thrips occurred in the tomato flowers.

Small plastic bags (5X8 cm) were used to enclose the thrips. Many tiny holes were punched in each bag with 000sized insect pins (BioQuip Products,Inc. 17803 La Salle Avenue, Gardena, Ca. 90248). The plastic bags containing the thrips were taken immediately to the greenhouse and placed around the flowers and fruit to avoid mortality observed when thrips remained for more than two hours inside the bag without the host. Each bag was sealed around the fruit or flower pedicel with parafilm paper (Parafilm "M" laboratory film, American National Can. Greenwich, Ct. 06830) and protected from direct sunlight with aluminum foil. Thrips were left inside the bag for five days; then, the plastic bags were removed and the thrips killed. Tomato fruits were harvested when physiologically mature but still green, and the number of scars on each fruit counted.

The female thrips used in these experiments were collected from wild and crop host plants. E. occidentalis were collected from flowers of strawberry, wild blackberries and








88

tomato. E. fusca were collected from peanut leaves and E. tritici from tomato flowers. They were taken into the laboratory for identification and counting using 2.6 to 40X magnification.

The relationship for each species between adult female density and fruit damage (number of scars per fruit) was determined by linear and curvilinear regression and significant differences (P<0.05) between individual treatments compared by a Duncan's (1955) new multiple range test. Also damaged small fruit were collected from tomato plants in production fields and the scars dissected and observed using 2.6 - 40X magnification in attempts to determine the cause of the damage. Others were placed in individual plastic vials and observed for up to 8 days to determine if larvae emerged.

Results and Discussion

Scarring on fruit in the checks was very low, but occurred on some dates because populations of these thrips sometimes occurred in the greenhouse (Table 5.1). However, the results of this study show that female thrips of E. occidentalis are responsible for the cosmetic fruit damage occurring in tomato in North Florida. The damage occured in this experiment when either fruit or flowers were infested. The number of scars per fruit caused by this species were significantly greater than the check, thereby demonstrating that the species does cause the scarring on tomato fruit. The mechanism by which the pest damaged the fruit was








89

investigated. Individual eggs and immatures were observed in the center of scars on damaged fruit. Immature thrips also occurred on damaged fruit, several days after adult thrips were removed from the plastic bags. These observations demonstrate that the damage -is caused by E. occidentalis oviposition into small fruit. Oviposition in small fruit and cosmetic damage were reported on grapes by Jensen (1973), Stafford (1974), and Yokoyama (1977) and on apples by Terry and DeGrandi-Hofman (1988).

Female E. fusca also were confined on flower and small fruit. The mean number of scars per fruit for all treatments were statistically similar to the controls. Female E. tritici were confined on flowers, but not small fruit (Table 5.1). As with E. fusca, there was no statistical evidence that the mature fruit had greater number of scars than the controls. Overall, these results indicate that E. fusca and E. tritici, although inhabitants of flowers, are not responsible for cosmetic damage on tomato fruit.

The relationship between the number of scars per tomato fruit and density per flower of E. occidentalis was evaluated by using regression (Fig. 5.1). The relationship was significantly quadratic (F=5.9, df=2, 146, p<0.01). The amount of variation explained by the quadratic model was very low (r2=0.07). Consequently, these findings indicate that the number of scars produced by individual females (or treatments of females) was very variable.








90
Oviposition by E. occidentalis, found to be the origin of the scars, is normally erratic and influenced by several factors. The, number of ovipositions is variable, ranging from 150 to 300 during a female life (Robb et al. 1987). The rate of oviposition is highly influenced by temperature (Lublinkhof and Foster 1977). They reported a mean of 24, 95, and 44 offspring per female at temperatures of 15, 20, and 300C, respectively. In this experiment, temperature inside the greenhouse was affected by the outside temperature which varied considerably between the period of March to July when the experiments were conducted. Another factor possibly affecting the results was the hosts from which the E. occidentalis thrips were collected. Female oviposition on flower structures other than the small, developing fruit would not be expected to result in fruit scarring and undoubtedly would increase variability of the results.

Although several species of thrips inhabit tomato flowers in North Florida, results from these experiments indicate that cosmetic fruit damage is caused by oviposition from E. occidentalis. The other species apparently are not economically important in causing cosmetic fruit damage. This explains why the damage was first noted in North Florida in 1985 (Olson and Funderburk 1986) shortly after E. occidentalis was first noted in the geographical area (Beshear 1983). Results also revealed that the relationship between the number of �. occidentalis female thrips in tomato flowers and the




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EISCPKI38_75PH4Q INGEST_TIME 2015-04-01T19:42:01Z PACKAGE AA00029879_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

SAMPLING PROGRAMS FOR AND ECOLOGICAL RELATIONSHIPS OF Frankliniella SPP. ( THYSANOPTERA : THRIPIDAE ) ON TOMATO. By VICTOR EBERTO ISALGUERO NAVAS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1990

PAGE 2

ACKNOWLEDGEMENTS I wish to express my sincere gratitude to Dr. Joseph E. Funderburk, Dr. Donald C. Herzog, Dr. Richard K. Sprenkel, and Dr. Steve Olson for serving on the advisory committee and for giving of their time and experience to improve my professional development in integrated pest management. Their constructive criticisms were invaluable in the course of this research and in the preparation of this manuscript. My gratitude is especially extended to Dr. Ramona Beshear, University of Georgia, for generously sharing her time and skills on the identification of thrips and to Dr. Timothy P. Mack, Auburn University, for his unconditional assistance in the statistical analysis of data. Sincere thanks are also due to the personnel of the North Florida Research and Education Center in Quincy and the Department of Entomology and Nematology in Gainesville. Especial thanks go to Myrna Litchfield, Sheila Eldrige, Tracey Austin, Jan Smith, Jan Gray, Janice Walden, Connie Rudd, Elizabeth Lewis, Andrew Brown, Glen Porcieau, and others for their endless patience and willingness to assist me during the development of this study. ii

PAGE 3

Thanks are extended to Dr. Avas Hamon and Dr. Charles Niblett for giving me advice and moral support whenever needed. Their warm friendship and exemplary manners will be always in my life. I wish to acknowledge to the Instituto de Ciencia Y Tecnologia Agricolas in Guatemala and its authorities for providing me with the scholarship to get my Ph.D. and especially for allowing me to properly orient my profesional development. To my loving daughter Cindy, my friend and son Juan Carlos, and my beautiful wife Lissette, I give thanks for their patience and confidence in my abilities. Their presence has made my hardships seem insignificant. Thanks my parents, Jose Victor and Maria Luisa, whose example of love and constant support were the basis for all I have achieved. They gave me more than I can ever repay. t • • 111

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii LIST OF TABLES vi LIST OF FIGURES vii ABSTRACT CHAPTER 1. INTRODUCTION 1 2. LITERATURE REVIEW 6 Insects and related pests in tomato 6 Flower thrips Frankliniella spp 8 General structure and systematics of the order Thysanoptera 8 Biology of Frankliniella spp 15 Ecology of Frankliniella spp 17 Host-plant relationships 17 Movement, migration, and dispersal 19 Survival and natural regulation 20 Economic importance 21 As pests on several crops 21 As predators of other pest species 24 Pollination of flower 24 Tomato spotted wilt virus (TSWV) 25 Host range of TSWV 26 TSWV transmission 26 Symptoms of TSWV 28 Management of Frankliniella spp. and TSWV 30 Pest management in tomatoes 30 Control of Frankliniella spp. and TSWV 33 Surveillance and sampling programs 38 Thrips sampling techniques 39 Spatial distribution patterns 43 Number of samples 46 Seasonal distribution 47 Binomial (or presence/ absence) sampling program 49 Factors affecting estimation of insect densities 51 Economic injury level and economic thresholds . 53 iv L

PAGE 5

3. SEASONAL PATTERNS OF Frankliniella SPP. ( THYSANOPTERA : THRIPIDAE ) IN TOMATO FLOWERS AND INFLUENCES OF SEVERAL FACTORS ON SAMPLE ESTIMATES 57 Introduction 57 Materials and Methods 58 Results and Discussion 61 4. BINOMIAL SAMPLING PROGRAM FOR Frankliniella SPP. IN TOMATO ACCORDING TO THE POSITION OF SAMPLING IN THE PLANT 73 Introduction 73 Materials ans Methods 75 Results and Discussion 77 5. FLOWER THRIPS, Frankliniella SPP. DAMAGE TO TOMATO THRIPS 85 Introduction 85 Materials and Methods 86 Results and Discussion 88 6. CONCLUSIONS 94 REFERENCES 102 BIBLIOGRAPHICAL SKETCH 114 v

PAGE 6

LIST OF TABLES Table 2.1. Table 2.2, Table 3.1, Table 4.1, Table 5.1, Key to the adult females of Frankliniella spp. thrips found in tomato flowers Developmental rate of Frankliniella spp. thrips Effects of and time of of adult F. triticiand in flowers during the seasons in field position, plant position, sampling on seasonal densities occidental is . F. fusca, and F. total adult and immature thrips in two tomato fields sampled spring 1988 and 1989 growing Gadsden Co. , Florida , Regression statistics of Taylor's power law relationships for Frankliniella spp. thrips sample data in tomato fields in North Florida during 1988 and 1989 The mean number of scars per mature green tomato fruit resulting from female Frankliniella spp. thrips confined on flowers and small fruit 12 18 72 82 92 vi

PAGE 7

LIST OF FIGURES Fig. 2.1. Key to the adult females of Frankliniella thrips found in tomato flowers 14 Fig. 3.1. Mean number of adult F. occidentalis . F. tritici and F. fusca per flower in tomato fields sampled weekly during the spring of 1987, 1988, and 1989 in Gadsden County, Florida 66 Fig. 3.2. Mean number of adult £. occidentalis and £. tritici per flower in tomato fields sampled weekly during the fall of 1987 and 1988 in Gadsden County, Florida 67 Fig. 3.3. Mean number of flowers and adult thrips per plant in the tomato fields sampled weekly during the springs of 1988 and 1989 and the fall of 1988 in Gadsden County, Florida 68 Fig. 3.4. Effect of field position of sampling on density estimates of Frankliniella spp. thrips in flowers in two tomato fields sampled during the spring in Gadsden County, Florida 69 Fig. 3.5. Effect of plant position of sampling on density estimates of Frankliniella spp. thrips in flowers in two tomato fields sampled during the spring in Gadsden County, Florida 70 Fig. 3.6. Effect of time of sampling on density estimates of Frankliniella spp. thrips in flowers in two tomato fields sampled during the spring in Gadsden County, Florida 71 Fig. 4.1 The relationships between the proportion of infested flowers and Frankliniella spp. density per tomato flower as estimated by using the Wilson and Room (1983) equation 83 vii

PAGE 8

Fig. 4.2. The relationships between the number of samples needed to estimate density at the 10 and 25% precision levels and the number of thrips per flower for Frankliniella spp. in tomato as determined by using the Ruesink (1980) equation 84 Fig. 5.1. Relationship between female F_. occidental is density and the number of scars per tomato fruit 93 • • • vm

PAGE 9

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SAMPLING PROGRAMS FOR AND ECOLOGICAL RELATIONSHIPS OF Frankliniella spp. ( THYS ANOPTERA : THRI PI DAE ) IN TOMATO By Victor Eberto Salguero Navas May 1990 Chairman: J. E. Funderburk Co-chairman: D. C. Herzog Major Department: Entomology and Nematology Flower thrips ( Frankliniella spp.) densities in tomato flowers were estimated weekly in commercial fields during the growing seasons of 1987, 1988, and 1989. Thrips were abundant during the spring but almost absent in the fall. F. occidental is (Pergande) , F. tritici (Fitch) , and F. fusca (Hinds) were commonly collected during the spring. Densities were greatest during May of each year. Species collected in the fall were F. tritici and F. occidental is . The effects of several factors on thrips density estimates were investigated. Plant position of sampling had a significant effect on density estimates of F. occidental is and F. tritici . but not F. fusca . Adult thrips of all species were more abundant in flowers located on the upper half of the plants than on flowers located on the lower half of the plants. Immatures were more abundant on the lower ix

PAGE 10

half. Sample location within a field did not affect density estimates of most thrips species. However, density estimates of F_. occidental is in 1988 were greater in tomato flowers located near margins than in flower located in nonmarginal areas. A similar tendency was observed in 1989 but this effect was not statistically significant. The time of day of sampling had no effect on thrips density estimates. Dispersion characteristics were guantified by using the Taylor's power law relationships for adult species and immatures. Most adult species and immatures were aggregated over a wide range of densities. F. fusca populations were slightly aggregated to random. A binomial sampling program was developed for adult F. occidental is . F. tritici . and F. fusca inhabiting the upper half of tomato plants. The slope and intercept of the Taylor's power law relationships were used to determine the relationships between proportion of infested flowers and density. The number of samples needed to estimate thrips densities at the 10 and 25% precision levels was also determined. Cosmetic fruit damage was found to be caused by oviposit ion from F. occidentalis . The other species were not found to cause cosmetic damage. The relationships between the number of female thrips in tomato flowers and the number of scars on the fruit was very variable. This is expected to hamper efforts to develop economic injury levels for £. occidentalis in tomatoes.

PAGE 11

CHAPTER 1 INTRODUCTION Flower thrips guild, Frankliniella spp. , constitutes one of the most complex polyphagous pest groups affecting the production of many ornamental and crop plant species. They are very general feeders (virtually cosmopolitan) on many plant species belonging to diverse families (Eddy and Livingstone 1931, Watts 1936, Newsom et al. 1953, Bryan and Smith 1956, Graves et al. 1987). The western flower thrips, Z. occidentalis (Pergande) , is considered an important pest in cotton (Stoltz and Stern 1978a, b, Rummel and Quisenberry 1979, Wilson 1982, Graves et al. 1987, Pickett et al 1988), table grapes (Yokoyama 1977), safflower (Carlson 1964b, 1966, Carlson and Witt 1977), onions (Harding 1961, Carlson 1964a, Dintenfass et al. 1987), many ornamentals (Baker and Stephan 1986, Jones and Moyer 1986) , apples (Terry and DeGrandiHoffman 1988) , and tomato (Oetting 1985, Olson and Funderburk 1986, Cho et al. 1989). The tobacco thrips, F. fusca (Hinds) , has been reported attacking peanuts (Morgan et al. 1970, Tappan and Gorbet 1979, 1981, Lynch et al. 1984, Tappan 1986a, b) , and cotton (Watts 1934, 1937a, b, Watson 1965, Rummel and Quisenberry 1979). The flower thrips, F. tritici (Fitch), has been reported from 1

PAGE 12

commercial roses (Stannard 1968), soybean (Marston et al. 1979), cotton (Watts 1934, Watson 1965), and peanuts (Morgan et al. 1970). F. bispinosa (Morgan) (there is no accepted common name) has been reported from peanuts (Morgan et al. 1970) . Flower thrips are important pests in tomato due to direct damage to the fruit which results in superficial or cosmetic damage which is not acceptable in a fresh market product by consumers or by national federal standard. Additionally, F. occidental is and F. fusca are vectors of tomato spotted wilt virus (TSWV) which seriously affects production of tomato and other food and ornamental crops worldwide (Cho et al. 1989). In North America, TSWV has caused economic losses in Canada (Allen and Broadbent 1986) , Hawaii (Cho et al. 1986), and Louisiana (Greenough et al. 1985) . Tomato is economically the most important vegetable crop grown in Florida (Pohronezny et al. 1986). Almost 98% of Florida tomatoes are grown for fresh market sales. Of the $711 million annual total U.S. farm value for fresh market tomatoes, Florida accounted for 53% ($377 million) (Clough 1987) . The total value of this crop in Florida increased from $377 millions in 1987 to $603 millions in the season of 19881989 (Anonymous 1989) . This crop is characterized by having a few major key pests and many secondary or minor pests whose relative

PAGE 13

3 importance varies both within and among major production areas. These pests can cause an economic reduction in optimum yield or damage the fruit which in fresh market tomatoes reduces the effective marketable yield because even slightly damaged fruit are rendered unmarketable. Deformation of fruit may be produced by the feeding of caterpillars, aphids, or thrips (Lange and Bronson 1981) . Neither the flower thrips complex nor TSWV have been considered as important tomato pests in Florida in the past (Pohronezny et al. 1986) . F. tritici . F. fusca . F. bispinosa . and F. cephalica masori Wats were reported in Florida as early as 1913 (Watson 1922, 1923). F. occidental is is not considered to be native to the Southeast; its natural range in the U.S. is the West-North regions and Canada (Baker and Stephan 1986) . However, Watson (1923) reported this species in South Florida as early as 1910. The first published report of this species from the Southeast was from Georgia on cotton flowers in 1981 (Beshear 1983) . However, it had been collected on soybean in North Carolina in 1978 (Baker and Stephan 1986) . It was later found for the first time in cotton in Louisiana in 1985 (Graves et al. 1987). Direct damage of Frankliniella spp. to tomato fruit was first recognized in North Florida as an economic problem in 1985 about the time F. occidentalis was first recorded in that area (Funderburk 1988). In addition to £. occidentalis . f. fusca . F. tritici and F. bispinosa have been found to inhabit

PAGE 14

tomato blooms (Salguero and Funderburk 1989) . The cosmetic injury produced by flower thrips on tomato fruit has become so commonplace that rejections of fruit by regulatory authorities has already occurred (personal observations) . TSWV, on the other hand, was first detected and confirmed in tomato in North Florida in 1986 (Sprenkel 1988) . It was also found on peanuts in the same area in 1989 (Sprenkel 1989) . The incidence of this disease in tomato fields increased from less than 0.5% in 1988 (Sprenkel 1988) to 2% in 1989 (Helena Puche, unpublished data) . The presence of TSWV and Frankiniella spp. in North Florida is causing alarm among tomato producers causing them to increase both the dosages and freguency of insecticide applications. Insecticidal control measures usually suppress the adult thrips populations; however, they rapidly rebound following treatment because of migration into the field. This migratory characteristic, polyphagous behavior, the presence of several species in the area and the wide range of wild and crop hosts of TSWV (Cho et al. 1989) complicates the development of pest management programs. The high cost of production for fresh market tomatoes reduces the options of pest management tactics and results in chemical control used in a preventive manner being used as the predominant tactic to ensure guality of the fruit. As a new problem in the area, several characteristics of the vector-virus relationship need to be understood in order

PAGE 15

5 to develop an appropriate pest management strategy. A sampling program suitable for management and research purposes is a necessary first step to understanding the population dynamics and other biological characteristics of this complex problem. This study was oriented to 1. Determine the species composition and their seasonal abundance and dispersion characteristics of the most important thrips species inhabiting tomato flowers in North Florida. 2. Develop a sampling program for thrips on tomato suitable for scouting and research programs. 3 . Characterize damage to tomato fruit by Frankliniella spp . thrips . 4. Determine the relationship between density of Frankliniella spp. thrips and tomato fruit injury.

PAGE 16

6 CHAPTER 2 LITERATURE REVIEW Insects and Related Pests in Tomato Tomatoes can be grown for processing or for fresh market, the latter being the most important in Florida (Clough 1987) . Their production methods differ greatly; however, the range of potential pests is similar in all tomato crops grown in the same area and season under similar conditions. Tomatoes are hosts for many kinds of arthropods whose importance varies within and among major production areas. Even when a few species are considered major key pests, there are many minor or secondary pests that under specific situations can become major pests. Their damage can cause an economic decrease in optimum yield or economic loss due to reduction in quality of the fruit below fresh market grade or for processing. The most common pests that directly destroy or affect the marketable product in Florida are the corn earworm, Heliothis zea Boddie; the southern armyworm, Spodoptera eridania (Cramer) ; and the tomato pinworm, Keiferia lycopersicella (Walsisngham) . Pests that reduce yield are the granulate cutworm, Feltia subterranea (Fab.); the serpentine leaf miner complex, Liriomyza spp. ; and the tobacco hornworm, Manduca sexta (Joh.) . Additionally, other pests are important in other areas or can become important in Florida: stink bugs,

PAGE 17

7 lygus bugs, thrips, mole crickets, flea beetles, whiteflies, aphids and mites. Aphids, leaf hoppers, threehoppers , thrips, and whiteflies may transmit viruses or mycoplasmas to tomato (Cantelo and Webb 1980, Lange and Bronson 1981, Pena 1983, Pohronezny et al. 1986). Predators and parasites that attack pest species are frequently abundant, maintaining certain pests at low population levels. Among the most conspicuous are lady beetles, bigeyed bugs, Geocoris spp., minute pirate bugs, Orius spp. and parasitic Hymenoptera. Unnecessary application of pesticides can affect these natural enemies and thereby contribute to outbreaks of secondary pests. Present tomato pest management systems utilize many resources, including host plant resistance, cultural control, natural and applied biological control, and chemical control. This is especially true for the most important pests. However, chemicals are widely used in a preventive manner to control them and ensure quality of the product. The use of pesticides varies from year to year and from one area to another, and may be limited by residue tolerance regulations (Tingey 1979, Lange and Bronson 1981). Lange and Bronson (1981) consider that there is a definite trend away from the constant use of pesticides on tomato and recommend their use only when needed but in the context of a well-rounded holistic or integrated pest management approach to insect suppression. Research is needed

PAGE 18

8 on sampling methods for some of the important tomato pests and most of the secondary pests. Establishment of economic thresholds and economic injury levels are also needed. These areas of investigation are difficult because of differences in cultivars, geographic and climatic areas, use of the crop, and many other variables. Flower thrips Frankliniella spp. General Structure and Svstematics of the Order Thvsanoptera Thrips are minute and slender-bodied insects (0.5 14 mm in length) that may easily pass unnoticed. Their most striking feature is the wide marginal fringes on the long and slender, usually present, four wings which gives the order its scientific name. Nearly all true leaf and flower thrips have fully developed wings (macropterous) but apterous forms (brachypterous) are common in one or both sexes of many species as found by Eddy and Livingstone (1931) in f. fusca . The head bears a pair of usually 7 or 8 -segmented antennae inserted at the front between the compound eyes. Three ocelli are also present on top of the head of most adults. The mouth parts, considered a unique characteristic feature in thrips, are of the sucking type with a stout, conical, and assymetrical proboscis (mouth cone) protruding beneath the head and often appearing to originate near or between the base of the front legs. The labrum and clypeus form the front of the proboscis, the basal portions of the maxillae form the sides, and the labium forms the rear. Three flexible stylets,

PAGE 19

the left mandible (the right is rudimentary) and two maxillary stylets, lie within this cone. The first thoracic segment is freely movable (Lewis 1973, Borror et al. 1976). The order Thysanoptera is divided into two suborders: Terebrantia and Tubulifera. Terebrantia has the last abdominal segment more or less conical or rounded, the females usually have a well developed saw-like ovipositor, the wings lie parallel to each other, and the males are always smaller and usually paler in color than females. Tubulifera, on the other hand, has the last abdominal segment drawn out into a tube, the wings overlap at rest so that only one is completely visible, the females lack an ovipositor and the males are often stouter than females with enlarged forelegs (Lewis 1973, Borror et al. 1976, Ananthakrishnan 1979). The classification of families in Thysanoptera remains controversial among specialists; however, Priesner's scheme (cited by Stannard 1968, Lewis 1973, and Ananthakrishnan 1979) has been widely accepted. Priesner's classification involves five families, one of them (Phlaeothripidae) in the suborder Tubulifera, and the others (Aelothripidae, Merothripidae, Heterothripidae, and Thripidae) in Terebrantia. The great majority of thrips including most of the species of economic importance belong to the family Thripidae, which is widely distributed throughout the world. Frankliniella . included in the family Thripidae, is a widespread genus through the world containing more than one

PAGE 20

10 hundred, mostly flower living species (Palmer et al. 1989). Karny, cited by Sakimura and O'Neill (1979) , erected the genus Frankliniella in 1910 and published the first key to world species two years later. Bailey (1957) considers that this genus is one of the most difficult to classify in Thysanoptera . According to the same author nearly 150 species have been described or transferred in and out of the genus. For example F. occidentalis . F. fusca, and F. tritici were first included in the genus Euthrips (Morgan 1913) . Bailey (1940) considered Frankliniella spp. to be among the seven most injurious thrips groups of major importance in the United States where the genus includes the following species: £. bispinosa (Morgan), which is limited largely to Florida; F. fusca Hinds, found widespread throughout the continent; £. qossvpiana Hood restricted at that date to Arizona and California; F. insularis (Franklin), known from Florida, Texas and Arizona; F. moultoni Hood and F. occidentalis (Pergande) , found commonly in the far western states but reported in Florida in 1910 by Morgan (1913) ; F. tritici (Fitch), abundant in the central and eastern states; F_. vaccinii Morgan, limited to Maine and F. williamsi Hood, known from South Carolina and Florida. The apparent absence of some Frankliniella species in many states may be due to the lack of published data and the scarcity of collected material from these areas at that date.

PAGE 21

11 Even when a knowledge of the species present may be of no significant importance because control measures for one will be equally effective against others, under some specific circumstances it may be highly desirable to recognize the species involved. It is especially true if we want to correctly apply an integrated pest management approach and we know that usually there are important biological differences to consider. For example, F. fusca requires approximately five days longer to complete its development than does F. tritici (Watts 1936) . £. occidental is and £. fusca are vectors of tomato spotted wilt virus, while F. tritici is not (Sakimura 1953, 1962, 1963). Identification of thrips at the generic and specific levels is very difficult and usually requires examination of specimens properly mounted on a glass slide with a compound microscope. Keys for species identification in the genus Frankliniella have been developed by Watson (1923), Stannard (1968), and more recently by the Thysanoptera specialist R. Beshear (unpublished data) (Table 2.1 and Figure 2.1). The criteria most used by thysanopterists in Frankliniella include number, shape, and size of antennal segments and setal type of antennal sensoria, etc. (Kono and Papp 1977) .

PAGE 22

12 Table 2.1. Key to the females of Frankliniella spp. thrips found in tomato flowers.* 1. Macropterous ; abdomen either pale yellow with brown patches on the meson or entirely dark brown 2 Brachypterous ; color generally dark brown to lighter brown especially the thorax and head (in part) ... fusca 2. Pedicel of antennal segment III with a distinctly thickened middle ring wich in profile appears as angulations; setal comb on posterior margin of abdominal tergite VIII incomplete 4 Pedicel of antennal segment III straight or nealy straight along sides; setal comb on posterior margin of abdominal tergite VIII complete or incomplete 3 3. Setal comb on posterior margin of abdominal tergite VIII complete; anteromarginal and anteroangular setae on pronotum of similar length; postocular setae longer, as stout as the interocellar setae (Fig. 2.1). occidentalis

PAGE 23

13 Table 2.1 (cont.) Setal comb on posterior margin of abdominal tergite VIII incomplete; anteroangular setae usually longer than anteromarginal setae on pronotum; postocular setae shorter and much more slender than the interocellar pair (Fig. 2.1) fusca 4. Pedicel of 3rd antennal segment with distal and basal parts equally diverging (Fig. 2.1) tritici Pedicel of 3rd antennal segment strongly angulated 12-14 urn wide with the distal and basal parts of the angulation diverging inward; anteromarginal setae on the pronotum usually less than 2/3 the length of anteroangular setae (Fig. 2.1) bispinosa Pedicel of 3rd antennal segment has smaller angulation usually 7.0-10.5 urn wide, the distal surface of the angulation is flat but the basal part diverging: anteromarginal setae usually longer, 0.7-0.8 times longer than the anteroangular setae (from Miami area) Cephalica * With permission from R. Beshear (unpublished data)

PAGE 24

14

PAGE 25

15 Biology of Frankliniella spp. Sexual reproduction is prevalent among the Thysanoptera ; however, parthenogenesis occurs in many species. Females often predominate in field populations and in some species males are rare or unknown (Lewis 1973). Watts (1936) found that F. tritici females greatly outnumber the males in nature from early spring through most of the summer, but during the autumn the proportionate difference is not so great. The same author also noted that all the progeny from unfertilized females were males. This last observation was also made by Eddy and Livingstone (1931) and Bryan and Smith (1956) for F. fusca and F. occidental is . respectively. The metamorphosis of thrips is intermediate between simple and complete. It is considered simple because the immature stages are very similar to the adults and more than one preadult instar has external wings. This metamorphosis is also considered complete because at least some of the wing development occurs internally and there is a quiescent instar preceding the adult. In Frankliniella spp., as most of the Terebrantia, the third and fourth instars are inactive, do not feed, and have external wings (Borror et al. 1976). Because of the confusion surrounding the type of metamorphosis of this order, variations in terminology for the immature stages have resulted and are common in the literature. Customarily, thysanopterists have called the first two feeding instars

PAGE 26

16 larvae and the remaining quiescent, nonfeeding, preadult instar prepupa and pupa (Watts 1936, Stannard 1968, Lewis 1973, Nugaliyadde and Heinrichs 1984). These traditional terms have been objected to by some entomologists who have supported the use of the terms nymph and pseudopupae (Bryan and Smith 1956, Bailey 1957, Lublinkhof and Foster 1977, Ananthakrishnan 1979) . Frankliniella spp. , like other genera in the family Thripidae, are oviparous. F. occidentalis females begin oviposition within 72 hours after adult emergence and oviposition normally continues intermittently throughout remaining adult life (Bryan and Smith 1956). Watts (1936) observed that during the colder months the oviposition period can be drastically reduced in £. tritici . The total number of eggs laid by most female thrips ranges from about 30 to 300 (Lewis 1973). Robb et al. (1987) found that an adult F. occidentalis can deposit 150-300 eggs during its life. Watts (1934) reports an average of 55.5 and 41.5 eggs per female in F. fusca and F. tritici . respectively. The same author (Watts 1936) found later that F. tritici oviposited an average of 28.61 eggs per female with a maximum of 119. This large variation in the number of eggs laid by different females suggests the possibility that a larger average could occur in the field under favorable conditions. Lublinkhof and Foster (1977) showed that temperature can dramatically affect the reproductive rate of F. occidentalis females. They found a

PAGE 27

17 mean of 24, 95 and 44 offspring per female at temperatures of 15, 20 and 30°C, respectively. Most Terebrantia deposit their eggs singly in an incision made in the plant tissue by means of the saw-like ovipositor. Oviposition can occur within leaf sheaths, at the flower base, cotyledons, petals, sepals or glumes (Lewis 1973) . Seedling cotton was successfully used as oviposition media when rearing F. tritici (Watts 1936). Bryan and Smith (1956) used bean pods for oviposition of F. occidental is . This same species is known to oviposit in the young berry grapes (Yokoyama 1977) . Peanut leaves have been used as oviposition and rearing media of F. fusca (Kinzer 1968, Kinzer et al. 1972). The duration of the life cycle varies with the species and temperature (Table 2.2) Watts (1934) found that £. fusca has a longer life cycle and lives longer in the adult stage than F. tritici under identical conditions. A similar duration of the life cycle of F. fusca at 27° C is reported by Kinzer et al. (1972). Lublinkhof and Foster (1977) report similar data for F. occidental is . These authors also found a significant effect of temperature on developmental time. All life stages developed more rapidly as temperatures increased. Ecology of Frankliniella spp. Host-plant relationship s Flower thrips are very general feeders and are capable of reproducing on a wide range of plant species. Lists of

PAGE 28

18 CD U u 3 o w oo rg CP o rLO CN o 1 o 1 1 l — CP p^. 4J , — 1 3 ro 00 CN cp "0 ro H H CP CO u p, CO m CN rH (0 CO CP ro rH J M CN rH (N ro CP CP H to u iH • e u IT) o o 01 o rH CN ro 0] CO 4J C 0) T3 •H u u o CN CN ro CN CN O CN ro ro U H 4J P CN cp CN ro CN LD ro ro CN rCO U =1 o CN r» pcp u CD U CN u ro CP Eu rH T> C (0 rH CO cw ro o CP 4J X! rH CD M C H H 0) rH -U N X) 4J C CO •H 5 bC CN ro

PAGE 29

19 recorded hosts have been reported for f. occidental is (Bryan and Smith 1956, Yudin et al. 1986, Stewart et al. 1989), F. tritici (Watts 1936, Stewart et al. 1989), and £. fusca (Eddy and Livingstone 1931, Newsom et al. 1953). These lists include abundant and diverse host plant species belonging to several unrelated families and orders. Annual and perennial weeds as well as winter, summer and perennial crops, recorded as hosts, provide the flower thrips with an assortment of uninterrupted suitable hosts during the year. Movement, migration, and dispersal Thrips are considered weak flying insects even though flight is by far their most important natural form of dispersal. Regular dispersal can occur by self-directed flight or by passive movement by wind currents. Their finely fringed wings enable them to remain airborne long enough for the wind to blow them to great heights and for long distances (Stannard 1968, Lewis 1973). Despite their great dependence on wind for dispersal some species seem to be able to see, orientate towards, and land on a surface of their choice. Yudin et al. (1987) found that F. occidental is was highly attracted to white traps. Some species of Frankliniella are more attracted to some plant species, varieties, or cultivars (Watts 1936, Yudin et al. 1988, Stewart et al. 1989). Mass flights occur sporadically within the flight period and in many species are reputedly associated with thundery weather (Lewis 1973). Stannard (1968) suggested that

PAGE 30

20 migration of F. tritici over large continental land masses might be associated with frontal winds. Survival and natural regulation Overwintering of Frankliniella spp.has not been studied in detail. Lewis (1973) considers that only a few species of thrips do not overwinter. Watts (1936) found that winter conditions in South Carolina permit continuous development of immature stages of F. tritici . though in smaller numbers and at a slower rate than during the summer. He found second instar larvae on oats shortly after a low temperature of 12°F. F. occidentalis has also been found during the winter in California (Bryan and Smith 1956) and southern Texas (Stewart et al. 1989). Eddy and Livingstone (1931) reported that only F. fusca females survive the winter in South Carolina. Temperature and rainfall are the two most important weather factors affecting the number of thrips (Lewis 1973) . Bryan and Smith (1956) considered that the distribution of rainfall may be more important than the total amount. Heavy rains of short duration probably limit the increase in population by beating large numbers of adults and immature forms into the ground. Similar observations are made by Watts (1936) . However, there are insufficient data to support these observations. Besides hazards of climate and change in the environment, many predators and parasites regulate thrips populations. Thrips are eaten by many general predatory insect species in

PAGE 31

21 the orders Hymenoptera, Diptera, Coleoptera, Hemiptera, Neuroptera, and even in Thysanoptera . Mites in the families Laelaptidae, Cheyletidae, and Amystidae have also been reported attacking thrips. Additionally, thrips form part of the diet of some amphibians, reptiles and insectivorous birds or can be eaten accidentally. Despite their small size, the eggs and larvae of thrips are parasitized by minute wasps in the families Eulophidae, Trichogrammatidae, Mymaridae and Scelionidae. Nematodes infesting the larvae, pupa, and adult thrips have also been reported. Mites are also known to be ectoparasites of thrips. Fungi belonging to the genera Beauvaria . Cephalosporium . and Entomopthora also induce mortality in natural populations (Stannard 1968, Lewis 1973, Ananthakrishnan 1979) . This long list of natural enemies of thrips suggests the potential possibilities of biocontrol agents. However, few attempts to introduce or encourage natural enemies to control pest thrips species have been reported . Economic Importance As pests on several crops Several hundred species of thrips are considered pests in many crops. They can cause serious direct damage to crops by feeding, oviposition or as vectors of plant pathogens. Direct injury produced is often slight but occasionally it may be severe and result in serious losses when heavy infestations occur. Thrips feed on most plant parts, except roots, but

PAGE 32

22 feeding is usually concentrated on rapidly growing tissues such as young leaves, flowers and terminal buds. The typical injury to plant tissue consists of silvering, scarring and distortion of leaves and fruits caused by feeding, and subsequent later discoloration due to excrement and the growth of molds (Lewis 1973) . Oviposition in newly developing fruits can produce a halo spot consisting of a small dark scar surrounded by whitish tissue (Jensen 1973) . Beside causing mechanical injury, some thrips transmit plant pathogens including toxaemias caused by toxin in salivary secretions, bacterial and fungal diseases spread by mechanical contact and viruses transmitted during feeding (Stannard 1968, Lewis 1973, Ananthakrishnan 1980, Walkey 1985) . Bailey (1940) listed the 32 thrips groups most destructive to crops in the United States. He placed the Frankli niella group among 7 groups or species of major importance. F. occidentalis is considered an important pest in cotton (Stoltz and Stern 1978a, b, Rummel and Quisenberry 1979, Wilson 1982, Graves et al. 1987), table grapes (Jensen 1973, Jensen and Luvisi 1973, Stafford 1974, Yokoyama 1977), apples (Terry and DeGrandi -Hoffman 1988) , saf flower (Carlson 1964b, 1966, Carlson and Witt 1977), onions (Harding 1961, Carlson 1964a, Dintenfass et al. 1987), tomato (Oetting 1985, Olson and Funderburk 1986, Schuster and Price 1987, Cho et al. 1989), ornamentals (Baker and Stephan 1986, Zur Strassen 1986), lettuce (Yudin et al. 1987, 1988), and, because of its

PAGE 33

23 extensive host range, probably in other crops not yet reported . £. fusca has been reported attacking peanuts (Morgan et al. 1970, Tappan and Gorbet 1979, 1981 Lynch et al. 1984, Tappan 1986a, b) and cotton (Watts 1934, 1937a, b, Watson 1965, Rummel and Quisenberry 1979, Eddy and Livingstone 1931) . However, its status as an economic pest is not clear. Watson (1965) , studying the control of thrips (including F_. fusca and F. tritici) in seedling cotton, found that where thrips were sufficiently abundant to produce plant damage, thrips control resulted in more uniform and vigorous plants but did not influence the rate of crop maturity or the total yield of seed cotton. Similar results have been obtained in peanuts by Morgan et al. (1970). Additionally, f. occidental is and £. fusca are vectors for the tomato spotted wilt virus (Sakimura 1962, 1963, Paliwal 1976, 1979, Cho et al. 1984, 1989, Allen and Broadbent 1986, Greenough et al. 1985). F_. tritici has been reported in soybean (Marston et al. 1979, Irwin and Yeargan 1980), cotton (Watts 1934, 1936, Watson 1965, Rummel and Quisenberry 1979), peanuts (Morgan et al. 1970), and strawberry, roses, alfalfa, peas, peach, apricot, plum, etc., which are damaged to varying extent by this species (Watts 1936) . £. bispinosa has been reported in peanuts (Morgan et al. 1970).

PAGE 34

24 As predators of other pest species Predation on other small animals is a common phenomenon among thrips. Some species are entirely carnivorous, however, most predatory thrips are usually polyphagous including animal and plant materials in their diet. Predatory thrips species feed on small, soft-bodied insects including other thrips, aphid nymphs, psocids, scales, and especially on the eggs of mites and Lepidoptera (Lewis 1973) . Probably the best documented case of predation of the genus Frankliniella is in cotton and peanuts. F* occidental is is an omnivore, consequently an important early-season predator of Tetranychus spp., an important pest in cotton (Trichilo and Leigh 1986, Pickett et al. 1988, Gonzalez et al. 1982). Boykin et al. (1984) listed F. fusca and F. tritici as predators on eggs of the two spotted spider mite, T. urticae . Koch on peanut fields. F. occidental is is considered by Lewis (1973) an important predator on Aelothrips fasciatus and Tetranychus telarius (L.), and F. fusca on T. telarius. Despite the apparent effectiveness of Frankliniella spp. as predators on mites, its potential as a biological control agent may be limited because of its important potential as a pest of many crops . Pollination of flowers Another beneficial characteristic of thrips is their role in the pollination of flowers. Their individual contribution is probably small but compensated for by the great numbers

PAGE 35

25 that may be present in flowers. For example, F. occidentalis has been reported to Increase pollination on onions (Carlson 1964a) . Individual F. tritici captured while flying carried large quantities of pollen among flowers, thereby playing an important role on cross-pollination of several plant species (Lewis 1973) . The economic importance of Frankliniella spp. as predators and/or pollinators needs to be better analyzed and is probably insignificant. However, these characteristics must be considered when implementing management programs in an IPM approach . Tomato Spotted Wilt Virus (TSWV) Tomato spotted wilt virus is a serious disease affecting the production of many food and ornamental crops worldwide (Greenough et al. 1985, Allen and Broadbent 1986, McRitchie 1986, Cho et al. 1986, 1989). TSWV was first described in Australia by Brittlebank in 1919 (cited by Cho et al. 1989). It has since become an important problem in the United States and Canada (Cho et al. 1988). TSWV is unigue among plant viruses in that it is the only member of its group for which identification is based on particle morphology, has one of the broadest host ranges of any plant virus, is highly unstable in vitro, and is the only virus transmitted in a persistent manner by certain species of thrips (McRitchie 1986, Cho et al. 1989). Their large isometric particles are 85 nm in diameter and are enclosed

PAGE 36

26 within a lipoprotein envelop constituting about 20% by weight of virus particles (Bos 1983, Walkey 1985). Other chemical and physical characteristics of TSWV may be found in Ie (1970). Host Range of TSWV TSWV, like its vectors, has a very wide host range that includes at least 200 plant species. While dicotyledonous plants (at least 192 species in 33 families) seem to be preferred, monocotyledonous species may also be infected (8 species in 5 families) (Cho et al. 1989) . Economically important crops affected include tomato, potato, bell pepper, tobacco, lettuce, peanut, pineapple, papaya, cucumber, and several flowering ornamentals such as chrysanthemum, dahlia, gloxinia, and gerbera daisies (Ie 1970, Baker and Stephan 1986, McRitchie 1986, Cho et al. 1989). Several annual and perennial weed hosts may serve as additional virus reservoirs. Cho et al. (1986) found 44 plant species (mostly weeds), representing 16 families, infected with TSWV in Hawaii. The wide host range of TSWV ensures its perpetuation and facilitates its spread into crop fields throughout the year in temperate and subtropical regions. This characteristic, on the other hand, complicates the implementation of strategies to manage TSWV. TSWV Transmission Transmission of TSWV can occur through seed, by sap (in laboratory), or by vectors. When transmission occurs through seed, the virus can be apparently carried in the teste but

PAGE 37

27 not in the embryo. Infection by seed, however, is confused and rates of infection from one to 96% have been reported. Transmission by artificial inoculation of sap using abrasives is easy, especially when extracts are prepared in neutral buffer containing reducing agents (Ie 1970) . Transmission by vectors is probably the most important means. Only thrips are known to be vectors of TSWV, and TSWV is the only confirmed virus transmitted by thrips (Ananthakrishnan 1980, Walkey 1985) . Eight species of thrips belonging to the family Thripidae have been identified as vectors of TSWV: £. occidental is . F. fusca, F. schultzei (Trybom) , F. moultoni Hood, F. tenuicornis (Uzel) , Thrips tabaci Lindeman, J_. setosus Moultan, and Scirtothrips dorsalis Hood (Sakimura 1953, 1962, 1963, Paliwal 1979, Ananthakrishnan 1980, Amin et al. 1981, Allen and Broadbent 1986, McRitchie 1986, Cho et al. 1988, 1989). However, Paliwal (1976) found that J_. tabaci did not transmit the virus in Canada. TSWV is transmitted by thrips in a persistent and circulative manner. Viruses transmitted in this way are characterized by a long acquisition feeding time, a latent period (12 hours or more) after feeding before transmission can occur, vectors retain the ability to transmit it for at least a week or for the remainder of their life, and finally the virus is retained through the moult of the insect (transstadial transmission) (Walkey 1985) . TSWV is acquired only in the larval stage of thrips but may be transmitted

PAGE 38

28 throughout the adult stage. The acquisition feeding period is 2-3 days for 1. tabaci , E. occidental is . and £. fusca (Sakimura 1962, 1963) with a shortest reported period of 15 minutes for T. tabaci (Ie 1970) . A 4-to 10-day incubation (latent) period, depending on the vector species, occurs before the virus can be transmitted (Ie 1970) . Sakimura (1963) found that the average latent period for F. fusca was 9.3 days with a range of 4-12 days. Transmission efficiency was increased with acquisition feeding times from fifteen minutes to four days (Walkey 1985) . Vectors are maximally infective 22-30 days after acquisition but sometimes retain the virus for life (Ie 1970) . The general average in retention period observed in £. fusca and T_. tabaci was 15.4 and 11.9 days, respectively, with a wider range in F. fusca (1-43 days) (Sakimura 1963) . Transovarian passage of TSWV has not been reported, and it is speculated that vectors do not transmit the virus to their progeny (Ie 1970, Walkey 1985) . Males and females and macropterous and brachypterous forms of F. fusca did not differ significantly in their vectoring ability (Paliwal 1976) . Enzyme-linked immunosorbent assay (ELISA) can be used to detect TSWV in individual thrips (Cho et al. 1988) . Symptoms of TSWV Characterization of the symptomatology of TSWV is difficult because of its wide host range. Hosts may show differing symptoms because of specific responses of the

PAGE 39

29 affected plant species and in part due to the many strains of the virus (McRitchie 1986). Many minor variants (strains) have been isolated which produce symptoms differing in severity. The most stable and important of these are: strains TB (tip blight) , N (necrotic) , R (ring spot) , M (mild) , VM (very mild); strains A, B, CI, 02, 0, and E; the "vira-cabeca" , and the tomato tip blight strain (Ie 1970) . Symptoms of TSWV can include local necrotic lesions followed by systemic necrotic patterns, leaf deformation, local black spots, systemic mosaic pattern of yellow and dark green specks, local chlorotic spots with necrotic centers, etc. The list of symptoms goes on and on and can be easily confused with symptoms induced by other viruses, fungal or bacterial pathogens, or nutritional disorders (Ie 1970). Symptoms on foliage of infected tomato plants are characterized by thickening of veins, downward curling of leaves, bronzing, and ring spotting. The entire plant is often stunted. However, symptoms are most noticeable and diagnostic on immature tomato fruit. Green fruits show light green rings with raised centers, giving the fruit a lumpy appearance, and rendering them unmarketable (McRitchie 1986) . Varied and complex symptoms patterns can be found in ripe tomato going from almost unnoticeable decoloration of the normal red color of the fruit to very well marked mosaic of different tones of red, yellow, and green (personal observation) .

PAGE 40

30 Management of Frankliniella spp. and TSWV Pest Management in Tomatoes Pathogens, insects, nematodes, and weeds are of constant concern to tomato growers and in many cases limit production. These pests are especially important for fresh market tomato because high quality of the fruit is required. Tomato growers in Florida have adopted a complex horticultural scheme to deal with these pests. Tomatoes are grown on raised, plasticmulched beds fumigated with broadspectrum biocides before planting. Pesticides have been traditionally sprayed every 37 days to prevent foliar pest problems. Frequently, a single problem can require special treatment. Approximately 34 insecticide applications have been made on a single crop in an attempt to control leaf miners ( Liriomyza spp.) in South Florida. Up to 60 fungicide sprays have been applied on some winter crops to prevent diseases, especially bacterial spot ( Xanthomonas campestris pv. vesicatoria) and late blight ( Phy tophtora infestans (Mont.)), and still heavy losses have sometimes resulted. These examples give an idea of the amount of money spent by tomato growers when dealing with pest problems. Approximately 20% of the money invested in tomato production is spent for pesticide spray materials (Pohronezny et al. 1986) . Despite the predominance of chemical control being widely used in tomato pest management, other tactics have been developed. Lange and Bronson (1981) consider that, beside

PAGE 41

31 pesticides, present tomato pest management systems utilize many resources including host plant resistance, cultural, and biological controls. However, the same authors stated that many kinds of constraints against development and utilization of pest management programs exist, including fruit quality standards, pesticide use limitations, and research on improving management programs. Several adverse consequences have been associated with the constant use of pesticides, such as increased costs of production, development of resistance to the insecticides used, reduction of predator and parasite populations, difficulty in registering new products, environmental contamination, and social concerns. These factors have created a definite trend away from the constant use of pesticides in the most important tomato production areas in the United States (Lange and Bronson 1981, Anonymous 1985, Pohronezny et al. 1986). The complex nature of the tomato pests requires that researchers from different disciplines participate in the development and implementation of management methods. IPM programs, oriented primarily to use pesticides only when needed, are being used in many tomato producing areas such as Florida (Pohronezny et al. 1986), California (Anonymous 1985), and other states and countries around the world (Lange and Bronson 1981) .

PAGE 42

32 IPM, as defined by Bottrell (1979) , is the selection, integration, and implementation of pest control (tactics) based on predicted economic, ecological, and sociological consequences. Under this approach, pests are treated as part of a crop production system that includes also the physical and biological environment. The emphasis of IPM programs is on anticipating and preventing problems when possible. Four components are essential in any IPM program: accurate identification of pests, field monitoring, control action guidelines, and effective methods for prevention and control (including the correct use of appropriate pesticides when needed) (Anonymous 1985) . Some of these components such as pest identification and control methods are self explanatory. Monitoring, by routine field checks, provides the information necessary to evaluate pest problems and make management decisions. Control action guidelines are numerical thresholds which reflect the population level that will cause economic damage and provide a way to decide whether actions are needed to avoid eventual losses. Monitoring was considered by Pohronezny et al. (1986) the most important feature of a tomato IPM program started in Florida in 1976. Under this program, fields were systematically scouted twice-weekly and the resultant biological data used in farm management decisions. The basis for monitoring are given by the development of sampling programs which include several components: sampling technique,

PAGE 43

33 number of samples, sample unit, factors affecting density estimates, sample allocation, spatial and seasonal distribution, insect stage to sample, etc. (Ruesink 1980, Wilson 1985, Pedigo 1989) . These components will be discussed in section 4.3. Despite the great interest in implementing IPM programs to control pests in tomato, Lange and Bronson (1981) stated that the present state of this technique is just at the threshold of understanding the intricacies of this pest-crop complex which also includes economic and social characteristics. Differences in cultivars, geographical areas, use of the crop, costs of production, consumer requirements on quality, etc. are some of the limitations to the implementation of IPM programs in tomato. Control of Frankliniella s pp. and TSWV The tactics that can be used effectively in the control of plant virus diseases involve evasive measures by reducing sources of infection, limiting spread by vectors, and to minimize the effects of infection on yield. Sources of infection can be reduced by destruction of infected plants in the crop, in other crop species, or weeds known to be reservoir hosts of the virus; planting virus-free seeds or vegetative stocks, modifying planting and harvesting procedures and dates (breaking an infection cycle) . Control of vectors is probably the procedure most commonly used to avoid the disease. However, breeding for resistance or

PAGE 44

34 immunity to a virus provides the best single solution to the problem of virus disease (Matthews 1970) . Naturally, the combination of two or more of the measures above described will improve the prevention of virus diseases. It is also essential that all growers follow those measures because if one fails in doing it, more sources of infection may be available. Probably the most complete research program oriented to the integrated management of TSWV is the one being conducted in Hawaii by Cho et al. (1989). These scientists identified some useful tactics that combined can minimize TSWV disease occurrence. These tactics include crop rotation with nonsusceptible crops to reduce the buildup of inoculum sources, crop placement to avoid planting TSWV-susceptible crops adjacent to each other, control of alternate TSWV vector hosts, use of virus-free seedling, regular applications of insecticides, fallow field areas with high disease incidence, and soil fumigation to eliminate thrips associated with crop debris. However, their conclusions stated that these tactics are not totally effective if virus and vector occurrence is high and frequent throughout the area. Under these conditions they suggest not to continue planting susceptible crops. In another area of research, the same authors are developing resistant varieties of tomato, having developed at present several lines highly resistant to TSWV.

PAGE 45

35 The only available method to control TSWV in tomato is through the use of insecticides to reduce the thrips vectors. Control of adult thrips moving into tomato fields is effective in reducing primary spread of TSWV and direct cosmetic damage to the fruit due to oviposition, while control of immatures reduces secondary spread of the disease. Many pesticides have been shown experimentally to effectively control Frankliniella spp. populations in several crops. These insecticides include abamectin, acephate, aldicarb, azinphos-methyl, carbofuran, chlorpyrifos, cyfluthrin, cypermethrin, diazinon, endosulfan, fluvalinate, formetanate hydrochloride, metamidophos , methomyl, mevinphos, monocrotophos , naled, phorate, and phosphamidon (Carlson 1964a, Morgan et al. 1970, Carlson and Witt 1977, Tappan and Gorbet 1981, Lynch et al. 1984, Oetting 1985, Robb et al. 1987, Cho et al. 1989). This long list suggests that chemical control of Frankliniella spp. is feasible and resistance to insecticides can be avoided or minimized by properly rotating chemicals. However, chemical control of these species seem to be uneconomical in some crops. For example, despite effective reduction of Frankliniella spp. populations in peanuts, onions, and cotton, yields were not increased when control measures were used (Carlson 1964a, Watson 1965, Morgan et al. 1970, Tappan and Gorbet 1981, Lynch et al. 1984). Similar results have been observed when trying to control TSWV or cosmetic damage due to F. occidental is oviposition. Cho et al.

PAGE 46

36 (1989) found that none of six insecticides, which were effective to control the vector, suppressed TSWV disease occurrence in lettuce. Metamidophos has proven effective in reducing Frankliniella spp. populations in tomato fields in North Florida, but cosmetic damage to the fruit has remained high even when the freguency of pesticide application has been increased (personal observation) , probably because invading populations reinfest the field. The economics of chemical control of Frankliniella spp. in some crops such as cotton, peanuts, and onions remains in a state of confusion and most researchers agree that control with pesticides is uneconomical. The situation in tomato is different; its production could be probably impossible without the use of insecticides if TSWV and its vectors, which can additionally cause cosmetic damage, are present. Some herbicides have been reported to have certain insecticidal properties on thrips. Laster et al. (1984) found that the herbicides dinoseb and dinoseb + MSMA showed some insecticidal activity against Frankliniella spp. when applied as a standard weed control in cotton. Dinoseb has been discontinued but similar possibilities could be considered. Biological control is another alternative for managing Frankliniella spp. Predators and parasitoids usually kill large proportions of field populations of thrips (Lewis 1973) ; unfortunately, the constant use of insecticides reduces their densities. Stoltz and Stern (1978a) found that reduction of

PAGE 47

37 Geocoris pallens Stall from the multiple dimethoate and naledtoxaphene applications to control lygus bugs, allowed populations of its prey £. occidental is to increase significantly. Other predators on F. occidental is are the minute pirate bug, Orius tristicolor White (Stoltz and Stern 1978b) and 0. insidiosus (Say) (Lewis 1973). Other predators, parasites, and pathogens inducing mortality in natural populations of thrips are potential sources of biocontrol agents for Frankliniella spp. and are mentioned by Stannard (1968) > Lewis (1973), Ananthakrishnan (1979), and Young and Welbourn (1988). However, few attempts (if any) to introduce or encourage natural enemies to control these pest species have been reported . The development of resistant varieties to thrips damage could be another alternative of control. Unfortunately, breeding for resistance to these pests has received very little emphasis, probably because resistant varieties are considered of not much value for a cosmetic damage problem like that produced by thrips on tomato fruit. However, resistant varieties of table grapes to £. occidental is damage (similar to the damage found in tomato) were reported by Jensen (1973). He found that while some varieties are not affected, the greatest damage occurs with the variety "Italia" whose skin may split at the site of a halo spot during ripening and lead to bunch rot. other examples of resistant

PAGE 48

38 varieties to Frankliniella spp. occur on peanuts, where the Spanish, Valencia, and Virginia type peanut plants are resistant to F. fusca (Smith 1979) , and cotton, where glabrous genotypes showed some resistance to thrips (Rummel and Quisenberry 1979) . Insecticides seem to be an obligatory alternative tactic when economic loss due to reduced quality of fruit for fresh market or for processing can be expected, or if TSWV is present in the area. This is the case with Frankliniella spp. in tomato in North Florida. Under these conditions, the immediate focus of pest management has to be directed to develop a program for the more precise and judicious application of pesticides, while other alternatives are evaluated and implemented. Surveillance and Sampling Programs Acquiring quantitative and qualitative information about the agroecosystem is a preliminary and indispensable phase in IPM programs. Any action to be taken against a pest will depend on its presence, abundance, and potential threat, documented through pest surveillance. Pedigo (1989) defines pest surveillance as the watch kept on a pest for the purpose of decision making. The major objectives are detection of species present and determination of population density, dispersion, and dynamics. Population density estimates are obtained in quantitative surveys by sampling. Sampling requires that a representative part of the total population

PAGE 49

39 be taken and estimates be based on that part. Therefore, surveillance requires the availability of sampling programs. The development of a sampling program requires an understanding of its components including spatial distribution, sampling technique (how to collect information for a single sample), location of samples, number of samples, when to start sampling, frequency of sampling, what life stage of the pest to sample, etc. (Wilson 1985, Pedigo 1989). Development of efficient, reliable sampling programs for insect pest management requires considerable research by entomologists in order to generate, quantify and understand biological information of the pest-host relationship and the factor affecting this interaction. Thrips sampling techniques Techniques of sampling thrips, that give qualitative and quantitative estimates of the size and distribution of populations, are briefly described in this review. A more complete description can be found in Lewis (1973) , Southwood (1978), and Irwin and Yeargan (1980). Thrips can be sampled in three main environments; i.e., soil, vegetation, and air. The technique will vary for each habitat, for each species or group, and according to the objective of sampling. Populations in soil are usually taken with a 10-cm-diameter soil-corer. Litter can be sampled within a grid by scraping it to soil level and collected. Thrips can be extracted from samples of soil or litter by a dry funnel method in which the insects are

PAGE 50

40 stimulated to move from samples by heat applied from above by light bulbs or infra-red heaters. This method is more efficient for adults than for larvae, and useless for pupae. Flotation techniques can also be used to separate specimens from soil and litter. Populations that emerge from soil in spring can also be sampled with emergence cages placed on the ground . Thrips can be sampled from vegetation by removing infested specific parts of the plant (or the whole plant) and processing in the laboratory. Thrips can be separated from vegetation by washing, repellents, heated funnels, cooling, etc. Small structures of the plant (leaves, flowers, etc.) can be placed into 70% alcohol and thrips separated by hand. Estimates from standing vegetation can be taken with a sweep net, suction samples, beating and shaking, or directly counting the thrips on the plant part of interest. Aerial population can be sampled with suction traps (absolute estimates) or sticky traps and water traps (relative estimates) . Suction traps can be used to measure the daily periodicity of flight, the specific structure of faunas, or to relate numbers taking-off with weather conditions. Sticky traps are useful for detecting the presence of flying thrips and especially to record times of emergence and seasonal changes in activity. Some species of thrips are attracted to pale colors, especially white. The appropriate color for the species of interest must be determined, or black can be used

PAGE 51

41 to avoid differential attraction when studying the species composition of an habitat. Several of the above listed techniques have been used to estimate Frankliniella spp. populations in many crops. These can be divided into direct observation or collecting techniques and delayed collecting techniques. The direct count technique or visual inspection is that in which thrips are counted, from the plant or plant part of interest, directly in the field. This method was described efficient for Sericothrips variabilis (Beach) in soybean, but not for F. tritici (Irwin and Yeargan 1980) . Letourneau and Altieri (1983) found the visual inspection technique more effective than sticky traps, pan traps, and malaise traps when sampling F. occidental is in squash. Direct collecting techniques are those in which thrips are collected without bringing plants or plant parts to the laboratory. These techniques include shaking or striking, aerial suction, D-vac suction sticky or cup traps, malaise traps, irritation, etc. Shaking has been used to sample £. occidentalis in apple (Terry and DeGrandi-Hof fman 1988) and grape blossoms (Jensen 1973) and Frankliniella spp. from whole cotton plants (Watson 1965, Laster et al. 1984). Aerial suction traps were found to be effective in sampling £. occidentalis in grapes (Yokoyama 1977) . The D-Vac suction procedure has been used effectively for Frankliniella spp. in cotton (Stoltz and Stern 1978a, Marston et al. 1979).

PAGE 52

42 Sticky traps have been used for sampling F_. occidental is in onions (Harding 1961) and F. tritici in roses and peonies (Webb et al. 1970). A modified sticky trap, using styrofoam or plastic cups, was used by Yudin et al. (1987) for F. occidental is in lettuce. The same authors evaluated several colors and found that white traps caught significantly more thrips than 14 other colors tested including several tones of yellow, green, red, blue, etc. The cage-aerosol or irritation technique was used for sampling £. tritici in roses (Ota 1968) and in soybean (Marston et al. 1979). However, this procedure was less effective than other techniques in both cases. Delayed collecting methods are those in which the whole plant or plant parts are removed for later processing in the laboratory to separate the thrips. Whole plants have been used for sampling F. occidental is in onion (Dintenfass et al. 1987) , cotton (Pickett et al. 1988), lettuce and weeds (Yudin et al. 1988), and several crop and non-crop plant species (Stewart et al. 1989). Flowers have been collected to sample Frankliniella spp. in apples (Terry and Degrandi -Hoffman 1988) , roses (Henneberry et al. 1964, Ota 1968), peanuts (Tappan 1986a, b, Tappan and Gorbet 1979, 1981), tomato (unpublished data), and several weed species (Yudin et al. 1986) . Terminal buds have been used to sample F. tritici in soybean (Irwin et al. 1979), and F_. fusca in peanuts (Tappan 1986a, b, Tappan and Gorbet 1979, 1981, Lynch et al. 1984). Leaves have been used in cotton (Pickett et al. 1988), soybean

PAGE 53

43 (Irwin et al. 1979), weeds (Yudin et al. 1986), and pyrethrum (Bullock 1963) . Extraction of thrips from foliage is the next step in the delayed collecting technique. Several procedures can be used to separate thrips from foliage (Bullock 1963) , including washing in alchohol or detergent (Ota 1968, Henneberry et al. 1964, Tappan 1986a, b, Tappan and Gorbet 1979, 1981, Dintenfass et al. 1987, Yudin et al. 1988) shaking devices (Henneberry et al. 1964), berlese funnels (Stoltz and Stern 1978a, Stewart et al. 1989), turpentine-vapor chamber (Yudin et al. 1986), and extraction of thrips by hand using an stereoscope (unpublished data) . Spatial distribution patterns The determination of spatial distribution patterns is necessary in the development of sampling procedures for pest species and their natural enemies in crop systems. This structural component and description of the condition of a population, also called dispersion characteristics, refers to how a population occurs in space. Various spatial patterns, in space and/or time, can be displayed by insect populations, random, uniform, and clumped. These are described in Southwood (1978) , Ruesink and Kogan (1975) , Ruesink (1980) , and Wilson (1985) . The spatial distribution patterns of a species affects the sampling program and the method of transforming the data prior to statistical analysis. Most insects rarely disperse themselves randomly in their natural

PAGE 54

44 environments. As a result their dispersion patterns do not follow a normal frequency distribution (Waters 1959) . However, most statistical methods are applied to data with an assumption of a normal frequency distribution. In order to overcome these problems, the actual data has to be transformed by a function whose distribution normalizes the data or stabilizes the variance. Southwood (1978) recommends the transformation of data from a clumped population by using logarithms, slightly clumped by using square roots, and uniform by using squares. A spatial pattern is random if every point on the surface has an equal probability of being occupied by an individual. In other words knowing the location of one other individual on the surface provides no information as to the location of any other individual. The poisson statistical distribution describes this spatial pattern. The clumped, aggregated, or contagious spatial pattern is described by the negative binomial distribution and occurs when the presence of an individual in a unit of habitat increases the probability of another occupying the same unit. This is the most frequent distribution encountered in insects (Southwood 1978) . Finally, the uniform or regular distribution occurs when the presence of an individual at one point decreases the probability of another being nearby. A species distribution pattern can be expressed in terms of a variance:mean ratio (s 2 /x) . When s 2 /x=l (or s 2 =x) , the

PAGE 55

45 distribution is considered random; if s 2 /x>l the distribution is considered clumped; and if s 2 /x
PAGE 56

46 (Elliot and Kieckhefer 1987, Funderburk and Mack 1989). It is also a useful tool in agricultural research for determining appropriate sample sizes for use in estimating the abundance of various species (Wilson 1985) and for developing binomial (or presence-absence) sampling programs (Wilson and Room 1983) . Number of samples The total number of samples for estimating insect densities will depend on the degree of precision desired or required. Optimal sample size determination allocates the minimal number of samples needed to achieve a certain level of precision. The reliability of the estimated density increases when sample size is increased; unfortunately, the cost of sampling is also increased incrementally. A proper balance between the reliability of the estimate and the cost of obtaining it must be estimated (Ruesink 1980, Pedigo 1989) . The degree of precision required depends on the purposes of the sampling program. Ruesink and Kogan (1975) distinguish between data used for insect pest management and data for research purposes. In the first case the idea is to know if the population exceeds a given threshold and considerable sampling error can be tolerated especially if the sample mean is exceptionally far from the economic threshold. In the second case emphasis is placed on the reliability of the parameter estimate and the sampling error must be equally small for all sample means.

PAGE 57

47 A series of equations for use in estimating sample sizes are described by several authors (Karadinos 1976, Southwood 1978, Ruesink 1980, Wilson 1985). Within a homogeneous habitat, Southwood (1978) considers that the number of samples (n) required is given by: n=(s/Ex) 2 ; where: s represents the standard deviation, x the mean and E the predetermined standard error as a decimal of the mean (i.e., normally 0.05). Pedigo (1989) suggests the formula n=(ts/Ex) 2 , where the t value is found usually at 0.05 probability. Wilson (1985) presents a generalized form which can be used for organisms having distribution patterns ranging from highly clumped to uniform: n=t 2 /2D" 2 .a.x b " 2 ; where t 2 /2 is the standard normal variate for a two tailed confidence interval, D is a fixed proportion of the mean and is used to define 1/2 of the confidence interval (1/2 C.I. = DX) , and a and b are Taylor's coefficients. Seasonal distribution Biological environments are characterized by having annual cycles of resources and unfavorable conditions. Ecologically, insects have developed a phenological "strategy" to adapt their life cycles to these environmental cycles. An understanding of this seasonality relationship is required to develop efficient pest management programs. A knowledge of the seasonal distribution of the dominant pest and beneficial species in the agroecosystem will determine, if necessary, when to apply appropriate control

PAGE 58

48 tactics. The seasonal distribution of some insect species varies according to several factors such as weather, geographical areas, crop, season, etc. However, other insect species follow typical seasonal distribution patterns among years, some examples are the pea aphid Acyrthosiphon pi sum (Maiteki et al. 1986), Nabis spp. (Shepard et al. 1974) and Geocoris spp. (Funderburk and Mack 1987) in soybean. Unfortunately, not all the species follow a typical predictable distribution pattern (earner et al. 1974). Harding (1961) , after studying thrips infestations in onions in South Texas, concluded that: a) no concentrated movement or thrips into onion fields occurred, b) thrips movement into onion fields lessened as other host plants became abundant and the onions approached maturity, c) precipitation and mean daily temperatures below 50 degrees F. reduced thrips movement, and d) destructive infestations resulted from build up by breeding in the field and not from thrips movement from outside of the field into the field. Race (1965) found a direct relationship between populations of F. occidental is before and after cotton plantings. With this data, post-planting populations could be predicted with accuracy. By observing the seasonal abundance records of F. occidentalis . He concluded that there was not a typical pattern for the seasonal distribution of the populations during the five years of his study.

PAGE 59

49 Watts (1936) found that both adult and larval populations of F. tritici were small in early and mid-winter (December 20 to March 20) in South Carolina. With the beginning of spring (March 20 to June 20) the population ordinarily remained small and thrips activity minimal. High population densities were reached in the summer (June 20 to September 20) followed by a final gradual decline. The arrangement of combined crops in the field can also affect thrips populations. Letourneau and Altieri (1983) found that the population of F_. occidental is was much greater on squash leaves in monoculture than in tricultures with corn and cowpea intercropped with squash. The minute pirate bug, Orious tristicolor . a thrips predator, exhibited a more rapid colonization rate in tricultures than in monocultures. Binomial for presence/absence) sampling program The direct counting of abundant individual organisms, many of which are too small to be detected without the aid of a microscope, is usually a slow, arduous and costly process. Indirect estimates of insect population densities, which eliminate the necessity for counting every organism per sample and relies solely upon the presence or absence of individuals in sampling units, has been developed (Gerrard and Chiang 1970, Wilson and Gerrard 1971, Gerrard and Cook 1972). The basic idea is quite simple: if the frequency distribution of an organism can be identified, then mean density can be

PAGE 60

50 estimated from the proportion of samples attaining or exceeding some specified density. If the dispersion of the population in a particular habitat can be described by the negative binomial distribution with a known k, the probability of a particular mean population can be estimated by the formula: P y =k[ (1/1-P 1 ) 1/k -l] ; where P y is the mean population, P» the probability of the insect being present in the sampling unit determined on the basis of presence or absence sampling, and k the dispersion parameter, a measure of the amount of clumping in negative binomial distribution: K=(X 2 )/ (s 2 -X) . It is reliable only if the critical density levels are related to values of p» less than about 0.8; above this uncertainty associated with the predictions becomes too great (Wilson and Gerrard 1971, Southwood 1978) . Wilson and Room (1983) incorporated Taylor's power law into a mathematical function relating the proportion of sample units having organisms present to the variance and mean per unit. Based on this relationship the control status or density of the species can be assessed. Binomial sampling programs have been developed for small and usually highly aggregated organisms such as mites (Wilson et al. 1981, Zalom et al. 1984, Nyrop et al. 1989), aphids (Wilson et al. 1983), larvae of the European pine sawfly (Wilson and Gerrard 1971) , and corn root-worm egg populations (Gerrard and Chiang 1970) .

PAGE 61

51 Factors affecting estimation of insect densities Several variables that can affect the estimation of insect population are usually not considered as required components in sampling programs. These factors can cause disparity among observations of various researchers and can result in serious mistakes being made when sampling data is being taken for insect pest management purposes. Time of day of sampling, hedge effect, neighboring vegetation, wind speed and direction, height and/or position of the sample unit in the plant, and other factors, are variables that can easily induce variability in the results of sampling and should be also considered in the development of sampling programs. Hedge row and neighboring fields can affect the diversity of the fauna and the distribution of species in the crop field. Hedge row may have an especially rich flora of woodland shrubs and trees with various grasses and herbs or other crops which allow them to support more species of insects than neighbouring crops. Many species move between these two habitats especially from hedges to crops. Lewis (1969a) , studying insect communities on vegetation in a mixed hedge row and in neighbouring fields of pasture and field beans, found that diversity of species was always greater in the hedge row, but fluctuated more than in the beans and short pasture, as different flowers bloomed. The distribution of individuals for specific species within the field has also been shown to be affected by hedge

PAGE 62

52 row. Lewis (1969b) found that most species populations were concentrated near the hedge row. Species of some families such as Empididae, Miridae, Syrphidae and some Lepidoptera were concentrated on the hedge row. Limothrips sp. and other species were also affected by this factor in minor but significant extent. Both wind and neighboring flora were responsible for this distribution. Similar results are reported for the tomato pinworm, Keiferia lycopersicella (Walsingham) by Pena (1983) , and for Liriomyza trifolii in tomato (Price et al. 1981). Sampling time is another variable that can affect insect populations. This variable can be induced for several factors such as temperature, humidity, light, wind, and others. Tappan (1986a) , studying the relationship of sampling time to F. fusca numbers in peanut foliage buds and flowers, found no effect on number per bud, but the number per flower increased significantly between 0800 and 1000 hours with the largest number occurring between 1100 and 1200 hours. Watts (1936) found that the greatest activity of F. tritici in cotton seedlings occurred between 1000 and 1630 hours. He also found that activity declined under higher humidity and lower temperatures near nightfall. Under similar conditions the female, which is larger, was decidedly less active than the smaller and more slender male.

PAGE 63

53 Economic Injury Level and Economic Thresholds The economic injury level (EIL) was initially defined by Stern et al. (1959) as the level at which damage can no longer be tolerated and therefore the level at or before which it is desirable to initiate deliberate control activities. This concept was developed largely as a means for more rational use of insecticides. The same authors defined the economic threshold (ET) as the population density at which control measures should be applied to prevent an increasing pest population from reaching the EIL. Both concepts have been criticized by several authors and even other names have been proposed and/or used. Pedigo et al. (1986), for example, considers that if the economic injury level is defined as a population density rather than an injury level, the EIL should be actually called "critical population density". However, both concepts are widely accepted and remain as some of the most pervasive and influential elements in agricultural pest management . According to Pedigo et al. (1986), four primary components affect the EIL: market value, management cost, injury per insect unit density, and host damage per unit of injury. The mathematical relationship of these components have been widely defined. Economic damage, according to Stern et al. (1959), is expressed as a monetary value and occurs when:

PAGE 64

54 C(a)
PAGE 65

55 Economic thresholds for flower thrips vary greatly between crops and depending upon the presence of TSWV. Few thresholds have been experimentally determined. Others have been adopted in extension programs based on experience. A threshold of more than 100 thrips per bloom has been adopted for £. occidental is in cotton blooms in several states to prevent poor pollination, blossom drop and small boll shed (Sprenkel 1987, Graves et al. 1987). However, a much lower threshold (2 thrips per plant) is used for the same species in the same crop but at seedling time (Race 1961) . Carlson (1966) experimentally determined an economic threshold of 10 20 adult F. occidental is per bud on saf flower plants. An economic threshold of 10 thrips per flower or florete (900 to 1000 flowers per head) was determined by Carlson (1964a) in onion seed plants. However, populations up to 3.6 thrips per flower increased pollination. Yokoyama (1977) found that clusters that supported up to 1582 adult and nymphs of £• occidentalis did not have a greater amount of surface scars than non infested clusters of seedless table grapes. "Detectable" levels of F. occidentalis constitute an economic threshold in ornamentals if TSWV is present but, if the disease is absent, thrips can be tolerated up to a level where premature aging of blossoms occur, depending on the plant species (Sprenkel 1987) . Populations of about 10 £. bispinosa thrips per flower increased bloom drop on tomato (Schuster and Price 1987). An average of 1 to 2 thrips per flower has been

PAGE 66

56 determined through experience to be the threshold for F. occidentalis in tomato for fresh market in the south east (Sprenkel 1987) . However, these numbers could be much lower if TSWV is present.

PAGE 67

CHAPTER 3 SEASONAL PATTERNS OF Frankliniella spp. ( TH YS ANOPTERA : THRI PI DAE ) IN TOMATO FLOWERS AND INFLUENCES OF SEVERAL FACTORS ON SAMPLE ESTIMATES Introduction Flower thrips, Frankliniella spp., are polyphagous insects that are pests of numerous crops worldwide (Cho et al. 1989). The western flower thrips, F. occidental is (Pergande) , and the tobacco thrips, F. fusca (Hinds), are pests of tomato because they vector tomato spotted wilt virus (Sakimura 1962, 1963). Adult female F. occidentalis oviposit in small tomato fruit, thereby causing cosmetic damage (unpublished data) . Although originally confined to the western half of the United States, £. occidentalis has moved into the Southeast, being first recorded in Georgia in 1981 (Beshear 1983). This species is now present throughout most of the region, including Florida (unpublished data) . Tomato spotted wilt virus has a wide host range and is worldwide in distribution (Cho et al. 1989), but has only recently been found in the Southeast. The disease was reported extensively in the region during 1986 on tomato, pepper, and other crops (Reddick et al. 1987, Hagan et al. 1987). It was first noted in Florida in 1986 (Olson and Funderburk 1986) . 57

PAGE 68

58 Efforts to develop strategies to manage thrips in tomato in the Southeast are hampered by a lack of important biological information. The species composition and seasonal abundances of thrips inhabiting tomato in the region are unknown. The primary purpose of this study was to determine seasonal abundance of thrips inhabiting tomato flowers. Multiple cropping seasons are typical of tomato production in the Southeast, with tomatoes in North Florida grown during the spring and fall. Consequently, seasonal abundance of thrips was determined for both cropping seasons. Another objective was to evaluate the effects on density estimates of sample location within a field, sample location on individual tomato plants, and the time of day when sampling. This information is needed to develop sampling programs to estimate density for scouting purposes. Materials and Methods Thrips densities in tomato flowers were estimated weekly in tomato fields during the spring growing seasons of 1987, 1988, and 1989 and during the fall growing seasons of 1987 and 1988. Two fields about 5 ha in size located in Gadsden County, Florida, were sampled. Each was a production field that was managed in a manner typical for the agroecosystem. Sampling in each field in all years began within one week of first blooming and continued until near final harvest. Flower samples for estimating thrips densities were placed in vials containing 70% ethyl alcohol and returned to

PAGE 69

59 the laboratory for further processing to determine the number of adults of each thrips species and the total number of adult and immature thrips. The contents of each vial was transfered to a 4-cm-diam petri dish and examined using 6.5 to 40X magnification. A microscope slide was prepared for each adult thrips, with CMC-10 (Masters Chemical Co., Inc., 520 Bonnie Lane, Elk Grove Village, Illinois 60007) used as the clearing medium. After at least 24 hr, adults were identified using 100 to 1000X magnification. Adult thrips were identified, by using a key to the genus Frankliniella developed by R. Beshear (unpublished data) (Table 2.1 and Fig. 2.1). During the spring and fall cropping seasons of 1987, the sampling protocol was designed with the sole objective of determining seasonal abundance. A total of 20 random samples of individual flowers were collected weekly in each field. During the spring and fall cropping seasons of 1988 and the spring cropping season of 1989, samples were collected to determine seasonal abundance and to evaluate the effects of several factors on density estimates. A total of 32 samples of individual flowers were collected weekly in each field, with the samples taken so that analysis of variance (ANOVA) procedures were used to evaluate the effects on density estimates of sample location within a field, sample location on individual plants, and the time of day when sampling. Consequently, the 32 samples taken in a field on each sample date were subdivided so that 4 random samples were taken from

PAGE 70

60 each of 8 treatments. Treatments were a factorial arrangement of two locations within the field (i.e., plants located < 40m from the field edges and plants > 40m from the field edges) , two locations on the plant (i.e., upper and lower half of each sampled tomato plant) , and two times during the day when taking samples (i.e., 0800 to .1000 hr and 1100 to 1300 hr) . Data from the spring growing seasons of 1988 and 1989 were used to evaluate these effects on density estimates of adults of each species and the total number of adults and immatures. Data for each growing season were analyzed as a splitplot randomized complete block over time, with treatments as whole plots and dates as subplots (Steel and Torrie 1960) . Individual fields were treated as block effects. The treatment sum of squares in each analysis was further divided into the main and interactive effects of sample location within fields, sample location on individual plants, and time of day when sampling. The mean number of adults of each species and of total adults and immatures per flower was determined for each sample date by averaging over all individual samples. The mean number of flowers per plant on each sample date also was estimated in each field during the 1988 spring and fall growing seasons and the 1989 spring growing season. For these estimates, the number of flowers on 6 randomly selected tomato plants was determined on each weekly sample date. The average number of thrips per plant was obtained then by multiplying the mean

PAGE 71

61 number of thrips per flower by the average number of flowers per plant. Results and Discussion Adult thrips accounted for about 88% of the total thrips (n=2353) collected in the tomato flowers during the three years of this study. About 97% of these adults were F_. occidentalis , F. tritici, and F_. fusca . Most of the remaining adults were F. bispinosa . with other species accounting for less than 1% of the total collected. These other species of adult thrips included Thrips t abaci Lindeman, Pseudothrips ineoualis (Beach) , Plesiothrips perolexus Beach, Chirothrips sp . , and Sericothrips sp . Adult thrips were present during each spring cropping season (Fig. 3.1) . Greatest densities occurred during May each year. Species included £. occidental is . £. tritici . and F. fusca. Populations of F. occidental is and F. tritici were abundant each year. Populations of £. fusca were present during two of the spring cropping seasons. Adult thrips were not present on most samples dates during the fall cropping seasons (Fig. 3.2). The species occurring in the fall included £. tritici and F. occidental is . Populations of F. tritici occurred each year. Populations of £• occidental is were present only during 1987. Adults of £. occidental is . £. tritici . and £. fusca rarely inhabited other tomato plant structures (unpublished data) . These species apparently prefer only the tomato flower,

PAGE 72

62 although insecticides used to control lepidopterous and other pests on tomatoes may have reduced any populations inhabiting other plant structures. Estimates of the number of flowers per plant on each sample date during the spring and fall growing seasons of 1988 and the spring growing season of 1989 were used to convert the number of adult F_. occidental is . F_. tritici. and £. fusca per flower to the number of thrips per plant (Fig. 3.3). For each species, seasonal patterns of abundance per flower were significantly (p < 0.001) correlated to seasonal patterns of abundance per plant. Correlation coefficients during the spring of 1988 for F. occidental is . F. tritici . and £. fusca were 0.97, 0.99, and 0.97, respectively. The correlation coefficient during the fall of 1988 for F. tritici was 0.99. Correlation coefficients during the spring of 1989 for jj\ occidental is . £. tritici . and Z^ fusca were 0.96, 0.98, and 0.98, respectively. The seasonal abundance data shows that thrips populations are more abundant at certain times of the year, with population trends apparently unrelated to crop phenology or the number of flowers per tomato plant. Populations of £. occidental is, F_. tritici . and F_. fusca were common between late April and early June, with greatest densities during May. Densities were very low on the other sample dates, especially during each fall cropping season. For each species, seasonal patterns of density per flower were very similar to seasonal patterns of abundance per plant.

PAGE 73

63 The effects of field position, plant position, and time of sampling on density estimates were determined using data from the spring growing seasons of 1988 and 1989 (Table 3.1). Seasonal densities of all species were low during the fall growing season of 1989, thereby eliminating the possibility of examining the effects of the above treatments on density estimates during that growing season. Plant position of sampling had a statistically significant effect on density estimates of adult F_. occidental is . adult f. tritici, total adults, and immatures (Fig. 3.5). Densities of these adult species were greater in flowers located on the upper half of the plants than in flowers on the lower half of the plants. Adult thrips, therefore, prefer the flowers on the upper half of the plants. Conversely, density estimates of immature thrips were greater for flowers located on the lower rather than upper half of the plants. Because of the very low numbers of immature thrips in all flowers, it is believed that insecticides sprayed in the fields greatly reduced their populations. There may have been better immature survival on the lower flowers, because of poorer coverage of insecticides. Sample location within a field significantly affected density estimates of F. occidental is during 1989, with greater populations near field margins (Table 3.1, Fig 3.4). This difference was not significant in 1988, but a singnificant field position * plant position and field position * time of sampling interactions revealed that field position affected

PAGE 74

64 density estimates. According to these interactions, density estimates on the upper half of tomato plants were greater on field margins than in the center of fields. Density estimates in marginal areas were greater between 1100-1300 hours than at 0800-1000 hours. Other adult species or immatures were not affected by location "of samples within a field (Fig. 3.4). Time of sampling had no significant effect on density estimates of adults of any species, total adults, or total immatures (Fig. 3.6). The results from these studies provide important information for integrated pest management programs. Economically damaging populations of thrips species capable of transmiting TSWV or causing cosmetic fruit damage occur most frequently during the spring cropping season. In North Florida, this period occurs between late April and early June. It should be possible to focus detection and management efforts to this period of the spring growing season. This data indicates that estimates of thrips densities in the flower also can be used to estimate thrips densities per plant. The three common species of thrips in tomato apparently are primarily flower inhabiting. However, numerous pests reach outbreak densities frequently in tomatoes, and insecticides are required as an obligate control measure. Insecticides sprayed in the commercial fields sampled in this study may have controlled any populations of thrips inhabiting other plant structures but not adult thrips in the flowers.

PAGE 75

65 The results involving the effects of several factors on density estimates of thrips provide important information when monitoring thrips in tomatoes. The time of day when sampling had no significant influence on density estimates of any species. Field position affected densities of F_. occidental is only. The location of flowers on plants greatly affected densities of F. occidentalis and F_. tritici . but not £. fusca . These effects on density estimates of some species will need to be considered in scouting programs. Separate estimates of density of F. occidentalis will need to be made for marginal and non-marginal areas of tomato fields. Also separate estimates of density of F. occidentalis and F. tritici will be needed for flowers located on the upper or lower half of tomato plants.

PAGE 76

66 3 2 1 F. occidentalis 1937 F. tritici F. fusca April ' May ' June ' July Fig. 3.1. Mean number of adult F_. occidentalis . F. tritici . and F. fusca per flower in tomato fields sampled weekly during the spring of 1987, 1988, and 1989 in Gadsden County, Florida.

PAGE 77

67 F. tritici 1987 F. occidentalis 0 1988 Aug 1Sept Oct Nov Fig. 3.2. Mean number of adult F. occidentalis and F. tritici per flower in tomato fields sampled weekly during the fall of 1987 and 1988 in Gadsden County, Florida.

PAGE 78

68 150 100 50 AUG ' SEPT OCT ' NOV 150 100 50 APRIL MAY JUNE JULY 150 100 50 APRIL MAY JUNE "JULY Fig. 3.3. Mean number of flowers and adult thrips per plant in the tomato fields sampled weekly during the springs of 1988 and 1989 and the fall of 1988 in Gadsden County, Florida.

PAGE 79

69 0) M C (0 u fa o 0) +j P 3 ^ u « C «M •P T3 M 2 «P . a re w B o *2 «w c u) c 5 O o VI H m >> H (0 c o> o •H «J p (0 w o o o o o c o •H P T3 O *4 -H S» u u ^ 0 O °«H fa H _ 5-H P H c 2 3 * a O +J rH •P P 0) -Q O T3 «J 0 . w XI M ftTJ O w a o a (0 O o •H fa

PAGE 80

70 (0 A C P •H P c o> id p Mti 3 5 25 S P -rt " H S»s & c w S o 2 p * ? O .5 C 4J H «j w P S | M 0) > •H P O 0) i <«P* g-H . +> Tl M TJ H P ^.H'H H 0) u jQ 8 0 H 2 DiH &, 0 , in •H

PAGE 81

71 w a >i Is p o ^ o ci c a a) i° H &-< to ssg» Ml -H O 10 O SI * J •w 2 a o i * o p 2 9 g c «J -H -H O jj a >i n W -P
PAGE 82

-H -P tr> B) H C C 3-H o -d in to o to tn C •H > O r-l CP oj oo 3 H p (0 T3 g corn 00 C T3 CO a) c ca .p •H ro 0) m O rH ro (N CO in rvo r~ an -I rH rH o o o o Si roo VO rH ro vo r— cm ro rH cm cn t in o o o o o o vo o o ro vo in cn o o o o r-. in ro CO VO T VO CO VO CM VO oo cm co oo vo co o in i — ro o o in in o o o o in rH o o I co ro vo ro O O rH rH HH ro ro O o rH r— oo rr— ^ m in B u MH ~ C « C -H E -H Cr>o Cr u u o o o o o o o o o ro o ro rH rH rH rH XX 11 o o o o o o o o 00 rH CO rH Q, 10 O 10 3£ J J3 6 o u vt-t — c 0J E-H 4-> o cn C "aui m a g ' — I > vw 9 U5 rH * ro in rH ro cm t in o o co oo vo • •••••• o cm o o ro o o O rH CM VO oo in O (N VO (N rCM O O rH o ro ro o in in cm in (N i — in VO rH rH fS O O in in o o o o o vo (N oo cn co o ro (N H CM ro rro o ro O rH o ro o o * CO VO CM CO in rH in ro o ro ro cm T O CM rH O O O O VO * t cm in o rH m o OHHNHNO O CM O O O O O VO rH VO O CO ^3* VO rro ro oo cn r~ CN O CN O rH CN rH VO r~ rH O VO rH CO
PAGE 83

CHAPTER 4 BINOMIAL SAMPLING PROGRAM FOR Frankliniella spp. IN TOMATO ACCORDING TO THE POSITION OF SAMPLING IN THE PLANT Introduction Several species of flower thrips, Frankliniella spp., are known to inhabit tomato plants and have been associated with cosmetic fruit damage (unpublished data) and/or transmission of tomato spotted wilt virus (TSWV) (Sakimura 1962, 1963). Direct damage to tomato fruit was first recognized in North Florida as an economic problem in 1985 about the time £. occidental is was first recorded in the tomato crop (Olson and Funderburk 1986, Funderburk 1988) . F. tritici, £. fusca . and F_. bispinosa have been also found to inhabit tomato flowers (Salguero and Funderburk 1989) . The cosmetic damage produced by flower thrips on tomato fruit has resulted in rejections of fruit or lowering of grade by regulatory authorities (personal observations) . TSWV is a serious disease affecting the production of tomato and other food and ornamental crops worldwide (Cho et al. 1989). Economic losses in tomato due to TSWV have been reported in Canada (Allen and Broadbent 1986) , Hawaii (Cho et al. 1986), and Louisiana (Greenough et al. 1985). This disease was first detected and confirmed in tomato in North Florida 73

PAGE 84

74 in 1986 (Sprenkel 1988) . The presence of Frankliniella spp. and TSWV in this region has resulted in an increase in both dosages and frequency of insecticide applications. The focus of integrated pest management programs is to estimate population densities of individual species and apply control tactics only if economic thresholds are reached. However, the direct counting of abundant and small individual organisms is usually a slow, arduous, and costly process. Binomial sampling is a practical and reliable procedure which eliminates the necessity for counting every organism per sample unit and relies solely upon the presence or absence of individuals in each unit. This technique is based upon the relationship between the proportion of sample units with one or more organisms (incidence) and the density of insects (mean) per sample unit. A sample estimate of incidence is then used to predict density (Gerrard and Chiang 1970, Wilson and Gerrard 1971, Gerrard and Cook 1972) . Sampling programs to practically and reliably estimate densities of flower thrips in tomato have not been developed and are needed to effectively manage these pests. Direct counting of flower thrips is technically difficult and a slow and arduous process, because of their small size and their tendency to quickly escape when disturbed. However, Frankliniella spp. thrips presence or absence in flowers apparently can be determined directly in the field. Consequently, binomial sampling may be a feasible field-

PAGE 85

75 sampling method for estimating Frankliniella spp. densities in tomato fields. The purpose of this study was to develop a sampling technique suitable for scouting programs to estimate density of flower thrips in tomato flowers. Dispersion characteristics of Frankliniella spp. were quantified and the data used to develop a binomial sampling program. Materials and Methods Population densities of flower thrips were estimated weekly in two tomato fields during the spring growing seasons of 1988 and 1989 and during the fall growing season of 1988. Each field was about 5 ha. in size and located in Gadsden County, Florida. Each field was managed in a manner typical for tomato production fields. Sampling began when first flowers became available and continued until near final harvest. A total of 13 and 12 samples were taken in each field during the spring and fall growing season, respectively. Each sample consisted of 32 sample units (individual open flowers) . Flowers were placed in individual vials containing 70% ethyl alcohol and returned to the laboratory for processing. The content of each vial was transferred to a 4-cm-diam petri dish and examined using 6.5 to 40X magnification. A microscope slide was prepared for each adult thrips, with CMC-10 (Master Chemical Co., Inc., 520 Bonnie Lane, Elk Grove Village, Illinois 60007) used as the clearing and preserving medium. Adults were identified 24 hours later using 100 to 1000X

PAGE 86

76 magnification. Adults were identified by using a key to the genus Frankliniella (Table 2.1, Fig. 2.1) developed by R. Beshear (unpublished data)'. The number of adults of each thrips species and the total number of adults and total number of immatures in each sample were determined. The 32 sample units (individual blooms) , taken in a field on each sample date, were subdivided into 8 treatments with 4 replicates for each treatment. Treatments were a factorial arrangement of the following three factors: two locations of sampling within the field (i.e.; margin, plants located < 40m from the field edges; and center, plants located > 40m from the field edges), two position of sampling on the plant (i.e.; upper and lower half of each tomato plant) , and two times during the day when taking samples (i.e.; 0800 to 1000 and 1100 to 1300 hr) . Plant position of sampling was the only factor which significantly affected density estimates of F. occidental is and F. tritici during each year (refer to Chapter 3). No factor significantly affected density of F. fusca. Taylor's power law relationships (Taylor et al. 1978) for each species therefore were determined for the upper half and the lower half of tomato plants and for data pooled over both plant positions. The same relationships also were used to describe dispersion characteristics of total adults and total immatures .

PAGE 87

77 The Taylor's power law relationships were used to develop the binomial sampling program for adult F_. occidental is . F_. tritici, and F. fusca inhabiting the upper half of tomato plants. The slope and intercept of the Taylor's power law relationships were used in the Wilson and Room (1983) equation to determine the relationships between mean density and the proportion of infested tomato flowers. The number of samples needed to estimate thrips densities at the 10 and 25% precision levels was determined by using the equation in Ruesink (1980) . Results and Discussion Most of the thrips collected in the tomato flowers through both growing seasons were adults, with only 12% being immatures. Three species accounted for about 97% of the adults collected, including £. occidentalis . F. tritici . and F_. fusca (Refer to Chapter 3) . Plant position of sampling significantly affected density estimates of adult F. occidentalis and £. tritici. with densities much greater on flowers located in the upper half of the tomato plants than on flowers located on the lower half of the tomato plants. Consequently, regression statistics of Taylor's power law relationships for individual sample estimates of F. occidentalis . F. tritici . and F. fusca adults and of total adults and total immatures according to plant position of sampling are given in Table 4.1. Taylor's power law relates variance (s 2 ) to mean density (m) by the relationship s 2 =am b . Taylor et al. (1978) considered the slope

PAGE 88

78 (b) to be a constant for a species (with values of b_ < 1, b = 1, and b > 1 indicating uniform, random, and aggregated distributions, respectively) and the intercept (a) to be reflected by sample unit size. Taylor's power law allows for a description of a species distribution pattern in relation to density. For nearly all of the thrips regression relationships, b was statistically > 1 (P <0.001) for a t test. These values of b indicate that F. occidental is . F. tritici . and total adults and immatures thrips were aggregated over a wide range of densities in tomato flowers. £. fusca showed an aggregated dispersion in 1988 (b = 1.66), but its dispersion in 1989 was random (b = 0.98). Apparently, F. fusca dispersion characteristics were slightly aggregated to random. Determination of the functional relationship between the proportion of infested tomato flowers and thrips density allows for estimation of thrips densities by presence or absence sampling. The functional relationships between the proportion of infested flowers and density were determined for F. occidental is. F. tritici . and F. fusca adults inhabiting flowers located on the upper half of the tomato plants. Estimating thrips densities in flowers on the upper half of the tomato plants is expected to be important for management purposes . For each flower thrips species, the relationship between the proportion of infested tomato flowers and density were

PAGE 89

79 similar at densities of about 1 thrips per flower (Figure 4.1). For example, a proportion of infested tomato flowers of 0.65, 0.66, and 0.67 represented a mean density of 1.1 adult thrips per flower for F_. tritici . F_. occidental is f and F_. fusca . respectively. Populations of E. occidental is and F_. tritici in tomato flowers were very aggregated and densities of these species greater than 1.4 thrips per flower could not be reliably estimated by the proportion of infested tomato flowers. Because populations of F. fusca are less aggregated than populations of F. occidehtalis and F. tritici . the binomial sampling program could be used to estimate densities up to about 2 thrips per flower. Because adults of each species did not exhibit the same dispersion patterns, combining counts of all species in the flowers to estimate total density of thrips would not be statistically reliable. Further, each species differs in economic importance. F. occidental is and F. fusca transmit TSWV, but F. tritici does not (Sakimura 1962, 1963). Cosmetic fruit damage is caused by oviposition by F. occidental is . with £. tritici and F. fusca apparently not causing cosmetic damage (refer to Chapter 5) . For these reasons, estimates of total thrips densities in flowers has limited management value in situations where each species are present. The number of samples needed to estimate density of £. occidental is and £. fusca in tomato fields by using the binomial sampling program was determined for the 10 and 25%

PAGE 90

80 levels of precision (Figure 4.2) . The number of samples needed to estimate density of £. tritici could not be calculated using the Ruesink (1980) equation because b_ values were greater than 2. The number of samples needed to obtain a 10% level of precision was above 100 for densities lower than 1 thrips per flower for F. occidental is and F_. fusca . About 90 samples were needed for densities above 1 thrips per flower for F. occidental is . Fewer samples were needed at high densities for F. fusca . because its populations are less aggregated than F. occidentalis . The number of samples needed to estimate densities of F. occidentalis and F. fusca at the 25% level of precision was between 15 and 20 for densities above 1 thrips per flower. The number of samples needed to estimate densities of F. occidentalis at densities higher than 1 thrips per flower was about 15. The number of samples needed to estimate densities of F. fusca was lower. A 25% level of precision usually is considered acceptable for extensive sampling programs. To obtain this level of precision, fewer than 20 presence or absence samples are needed to estimate densities of F. occidentalis or F. fusca . Consequently, the binomial sampling program appears feasible to estimate individual densities of adult thrips species. However, the implementation of this sampling program in areas where two or more species of Frankliniella occur simultaneously probably will not be practical. Because adults

PAGE 91

81 of each species did not exhibit the same dispersion patterns, combining counts of all species and quantifying dispersion characteristics of all species is not statistically valid. Estimates of density of all species combined has limited management value anyway, because each species differs in economic importance. F_. occidental is and F. fusca transmit TSWV but F_. tritici does not (Sakimura 1962, 1963). Cosmetic fruit damage occurs when F. occidental is oviposits on small tomato fruit, with F. tritici and F. fusca apparently not responsible for cosmetic fruit damage (refer to Chapter 5) . Also, it is not possible to reliably identify Frankliniella spp. thrips in the field. Because several species of thrips commonly inhabit tomato flowers in North Florida, the use of a binomial sampling program will have little value for integrated pest management programs. However, the binomial sampling program will be useful for scouting programs in geographical areas where only one species commonly inhabit tomato flowers.

PAGE 92

82 0 c (0 Q) o i D O P rcn cn cn o o cm in r» co o o CM CM cm cn CM VO 0\ cn cn o o cn o cn o o H vo co vo cn h o o cn cn cn cm vo o cn cn co o o co H r» cn o o o o co cn co in o o VO CO cm cn CM CM CM o o h in in h cn cn cn o o co o CO CO CM o w
PAGE 93

83 I I I 2.0 3.0 No. thrips per flower Fig. 4.1. The relationships between the proportion of infested flowers and Frankliniell a spp. density per tomato flower as estimated by using the Wilson and Room (1983) eguation.

PAGE 94

84 200175 -• Precis on level 150 125 25 % 100 766025 No. thrips per flower Fig. 4.2. The relationships between the number of samples needed to estimate density at the 10 and 25% precision levels and the number of thrips per flower for Frankliniella spp. in tomato as determined by using the Ruesink (1980) equation

PAGE 95

CHAPTER 5 FLOWER THRIPS, Frankliniella spp. DAMAGE TO TOMATO FRUIT Introduction Flower thrips, Frankliniella spp., are important pests in agriculture because of the capacity of some species for transmiting tomato spotted wilt virus (TSWV) and the direct damage when feeding or ovipositing on foliage, flowers, or fruits. Because of the wide host range of Frankliniella spp. (and TSWV) , serious economic losses occur on cotton, peanuts, onions, and other crops in the southeastern U. S. (Sprenkel 1987) and on tomato in Canada and Hawaii (Allen and Broadbent 1986, Cho et al. 1989). However, flower thrips have not been considered important pests of tomato in the continental U. S. (Lange and Bronson 1981, Pohronezny et al. 1989). TSWV was only recently found in the southeastern U. S. on tomato, peppers, and other crops (Kucharek 1986, Olson and Funderburk 1986, Hagan et al. 1987, Reddick et al. 1987, Sprenkel 1988). Simultaneously, flower thrips were implicated as the cause for cosmetic fruit damage occurring on tomatoes grown in North Florida (Olson and Funderburk 1986, Funderburk 1988) . This cosmetic damage consisted of small dark scars or indentations usually surrounded by a lightened or whitish skin area. A similar damage, caused by oviposition of the western 85

PAGE 96

86 flower thrips, F_. occidentalis . was reported on grapes in California (Jensen 1973, Stafford 1974, Yokoyama 1977) and apples in Arizona (Terry and DeGrandi-Hof fman 1988) . Although tomatoes are California's most important vegetable crop (Anonymous 1985) and f . occidentalis is present and common, cosmetic damage to the fruit has not been reported. The same condition applies for Canada and Hawaii. The cosmetic fruit damage on tomato has become a serious economic problem in North Florida. Rejection or lowering of grade of damaged fruit has occurred and insecticide applications increased. An understanding of the cause of the damage on the tomato fruit and its association with Frankliniella spp. is necessary to implement management programs. The major purpose of this study was to determine if the Frankliniella spp. thrips inhabiting tomato flowers cause cosmetic damage to tomato fruit. The relationship between the number of scars per fruit and the thrips densities on tomato flowers and small fruit also was guantified. Materials and Methods Several densities of adult female flower thrips (0, 1, 2, 4, 8, and 16) of three species (F_. occidentalis . F_. fusca, and F. tritici ) were confined on individual small tomato fruit and flowers. Tomato plants for this experiment were grown in a greenhouse at about 27°C at the North Florida Research and Education Center, Quincy, Florida, from March to July 1989. Small fruits and open flowers were infested with F_.

PAGE 97

87 occidental is and F. fusca, while only open flowers were infested with F. tritici . More than 500 experimental units were established but about 25% of the infested flowers aborted (affecting all treatments) . This abortion occurred mainly during the last days of the study when temperatures were often greater than 27°C. Insecticides to control thrips were applied when natural infestations of thrips occurred in the tomato flowers. Small plastic bags (5X8 cm) were used to enclose the thrips. Many tiny holes were punched in each bag with 000sized insect pins (BioQuip Products , Inc . 17803 La Salle Avenue, Gardena, Ca. 90248). The plastic bags containing the thrips were taken immediately to the greenhouse and placed around the flowers and fruit to avoid mortality observed when thrips remained for more than two hours inside the bag without the host. Each bag was sealed around the fruit or flower pedicel with parafilm paper (Parafilm "M" laboratory film, American National Can. Greenwich, Ct. 06830) and protected from direct sunlight with aluminum foil. Thrips were left inside the bag for five days; then, the plastic bags were removed and the thrips killed. Tomato fruits were harvested when physiologically mature but still green, and the number of scars on each fruit counted. The female thrips used in these experiments were collected from wild and crop host plants. F. occidental is were collected from flowers of strawberry, wild blackberries and

PAGE 98

88 tomato. F. fusca were collected from peanut leaves and F. tritici from tomato flowers. They were taken into the laboratory for identification and counting using 2.6 to 40X magnification. The relationship for each species between adult female density and fruit damage (number of scars per fruit) was determined by linear and curvilinear regression and significant differences (P<0.05) between individual treatments compared by a Duncan's (1955) new multiple range test. Also damaged small fruit were collected from tomato plants in production fields and the scars dissected and observed using 2.6 40X magnification in attempts to determine the cause of the damage. Others were placed in individual plastic vials and observed for up to 8 days to determine if larvae emerged. Results and Discussion Scarring on fruit in the checks was very low, but occurred on some dates because populations of these thrips sometimes occurred in the greenhouse (Table 5.1). However, the results of this study show that female thrips of E. occidentalis are responsible for the cosmetic fruit damage occurring in tomato in North Florida. The damage occured in this experiment when either fruit or flowers were infested. The number of scars per fruit caused by this species were significantly greater than the check, thereby demonstrating that the species does cause the scarring on tomato fruit. The mechanism by which the pest damaged the fruit was

PAGE 99

investigated. Individual eggs and immatures were observed in the center of scars on damaged fruit. Immature thrips also occurred on damaged fruit, several days after adult thrips were removed from the plastic bags. These observations demonstrate that the damage is caused by F_ . occidental is oviposition into small fruit. Oviposit ion in small fruit and cosmetic damage were reported on grapes by Jensen (1973), Stafford (1974) , and Yokoyama (1977) and on apples by Terry and DeGrandi-Hofman (1988) . Female F. fusca also were confined on flower and small fruit. The mean number of scars per fruit for all treatments were statistically similar to the controls. Female F. tritici were confined on flowers, but not small fruit (Table 5.1). As with F. fusca, there was no statistical evidence that the mature fruit had greater number of scars than the controls. Overall, these results indicate that F. fusca and F. tritici . although inhabitants of flowers, are not responsible for cosmetic damage on tomato fruit. The relationship between the number of scars per tomato fruit and density per flower of F_. occidental is was evaluated by using regression (Fig. 5.1). The relationship was significantly quadratic (F=5.9, df=2, 146, p<0.01). The amount of variation explained by the guadratic model was very low (r 2 =0.07). Consequently, these findings indicate that the number of scars produced by individual females (or treatments of females) was very variable.

PAGE 100

90 Oviposition by £. occidental is . found to be the origin of the scars, is normally erratic and influenced by several factors. The number of ovipositions is variable, ranging from 150 to 300 during a female life (Robb et al. 1987). The rate of oviposition is highly influenced by temperature (Lublinkhof and Foster 1977). They reported a mean of 24, 95, and 44 offspring per female at temperatures of 15, 20, and 30°C, respectively. In this experiment, temperature inside the greenhouse was affected by the outside temperature which varied considerably between the period of March to July when the experiments were conducted. Another factor possibly affecting the results was the hosts from which the £. occidental is thrips were collected. Female oviposition on flower structures other than the small, developing fruit would not be expected to result in fruit scarring and undoubtedly would increase variability of the results. Although several species of thrips inhabit tomato flowers in North Florida, results from these experiments indicate that cosmetic fruit damage is caused by oviposition from £. occidental is. The other species apparently are not economically important in causing cosmetic fruit damage. This explains why the damage was first noted in North Florida in 1985 (Olson and Funderburk 1986) shortly after F. occidental is was first noted in the geographical area (Beshear 1983) . Results also revealed that the relationship between the number of £• occidental is female thrips in tomato flowers and the

PAGE 101

91 amount of damage was very variable . This is expected to hamper efforts to develop economic injury levels for F. occidentalis in tomatoes.

PAGE 102

•H O -P (0 e o 4J T3 C (0 o n s jC EH rH in Q) rH JQ CO EH (x,| (0 -u c o 1 Oi o u> in o o 1 o o 1 + i +[ +1 + i ro r<* o 1 CTi o CD in 1 o o o o 1 CN cn ro (N l rH rH crv 1 (0 ra c0 (0 o o o 1 CO vo o T 1 o o 1 o O 1 +1 + i +1 + 1 o o o l CN CN o 1 o o rH o H o rH 1 cr, in | CO +1 + 1 + 1 + 1 + l + i o o LT) CTi ID in CN CTi CO r— o in ro rH rH 1 CO CO iH rH rH rH ro rH Cr, ro H XI a a X ro a r« CN rH VO o rCT> rH V£> LO O rH rH o CN CN + 1 + 1 +1 +1 +1 + 1 Ol ro \o Ol o> i rH J-> u H IH •H •H W 4-> 0 a OJ u ro u rH -P 0 CJ QJ c •H ran 5 -H a> S rH a, 01 •H E 4-> rH 0)

PAGE 104

CHAPTER 6 CONCLUSIONS Tomatoes produced for fresh market sales is economically the most important vegetable crop grown in Florida. Many pest species infest and economically affect tomato production in the world. Several technigues to control these pests have been developed. An additional guild of pest species, flower thrips ( Frankliniella spp.), was detected in this crop in 1985 in North Florida. These species of thrips were related to cosmetic fruit damage and to the presence of TSWV. Each problem alone is a serious threat to tomato production. Both the cosmetic damage and TSWV have reached unacceptable levels and have reguired an increase in the use of pesticides in North Florida. Thrips are the only known vectors of TSWV (Walkey 1985) . Eight species of thrips belonging to the family Thripidae have been identified as vectors of TSWV: F_. occidental is . £. fusca, F_. schultzei (Trybom) , £. moultoni Hood, F. tenuicornis (Uzel) , Thrips tabaci Lindeman, T. setosus Moultan, and Scirtothrips dorsalis Hood (Sakimura 1953, 1962, 1963, Paliwal 1979, Ananthakrishnan 1980, Amin et al. 1981, Allen and Broadbent 1986, McRitchie 1986, Cho et al. 1988, 1989). f. tritici, another flower thrips found in tomato flowers, does 94

PAGE 105

95 not transmit the virus (Sakimura 1953) . The capacity of F_. bispinosa f also found in tomato flowers, as a vector of TSWV is unknown. As a new problem in the area, several characteristics of the vector-virus relationship need to be understood in order to develop an appropriate strategy to manage these pests. My studies were conducted to determine some ecological relationships between flower thrips and the tomato plant. The specific objectives were: 1) Determine the species compositon and the seasonal abundance and dispersion characteristics of the most important thrips species inhabiting tomato flowers in North Florida. 2) Develop a sampling program for thrips on tomato suitable for scouting and research programs. 3) Characterize damage to tomato fruit by Frankliniella spp. thrips. 4) Determine the relationship between density of Frankliniella spp. thrips and tomato fruit injury. The results from these studies provide important information for integrated pest management programs of tomatoes. Adult thrips accounted for about 88% of the total thrips collected (n=2353) in the tomato flowers during the three years of these studies. Three species accounted for 97% of all the adults collected. F. occidental is , present mostly during the spring, was the most abundant (53%) ; F. tritici, present in both, the spring and fall cropping seasons, accounted for 37% of the adults collected; and F. fusca,

PAGE 106

96 present only in the spring, accounted for only 7% of the adults. The seasonal distribution of thrips varied between seasons and within each season. Thrips were more abundant during the spring cropping season and present in very low numbers during the fall cropping season. In North Florida, the spring cropping season for tomato occurs between late April and early July. The tomato fall cropping season occurs between late August and early November. Thrips densities reached economically damaging levels during May of each year. These results indicate that field monitoring of thrips populations is necessary to orient farm management decisions. Proper estimates of thrips populations will suggest when to apply pesticides which is the only currently available option to control these tomato pests. The results involving the effects of several factors on density estimates of thrips provide important information for sampling programs. Plant position of sampling had a significant effect on density estimates of £. occidentalis and F. tritici . but not F. fusca. Adult thrips of the first two species were more abundant in flowers located on the upper half of the plants than in flowers located on the lower half of the plants. Sample location within a field affected densities of F. occidentalis only. This species was more abundant in tomato flowers located near margins than in flowers located in nonmarginal areas. The time of day of

PAGE 107

97 sampling had no effect on thrips density estimates. These differential effects on density estimates of some species will need to be considered in scouting programs. Separate estimates of density of £. occidentalis will need to be made for marginal and nonmarginal areas of tomato fields. Also separate estimates of density of £. occidentalis and F . tritici will be needed for flowers located on the upper or lower half of tomato plants. Dispersion characteristics of Frankliniella spp. were guatified and the data used to develop a binomial sampling program. Binomial sampling (i.e., presence: absence sampling) is a practical and reliable procedure which eliminates the necessity for counting every organism per sample unit and relies solely upon the presence or absence of individuals in each unit. This technigue is based upon the relationship between the proportion of sample units with one or more organisms (incidence) and the density of insects (mean) per sample unit. A sample estimate of incidence is then used to predict density (Gerrard and Chiang 1970, Wilson and Gerrard 1971, Gerrard and Cook 1972). Dispersion characteristics were guantified by using the Taylor's power law relationships (Taylor et al. 1978). For nearly all of the thrips regression relationships, b was statistically > l (p < 0.001) for a t test. These values of b indicate that £. occidentalis . F_. tritici . and total adults and total immature thrips were aggregated over a wide range

PAGE 108

98 of densities in tomato flowers. £. fusca dispersion characteristics were slightly aggregated to random. A binomial sampling program was developed for adult F_. occidentalis . F. fusca . and F_. tritici inhabiting the upper half of tomato plants. For each flower thrips species, the relationship between the proportion of infested tomato flowers and density were similar at densities of about one thrips per flower. A proportion of infested tomato flowers of 0.65, 0.66, and 0.67 represented a mean density of 1.1 adult thrips per flower of F. tritici . F. occidentalis . and £. fusca . respectively. The number of samples needed to estimate density of F_. occidentalis and F. fusca in tomato fields by using the binomial sampling program was determined for the 10 and 25 % levels of precision. The number of samples needed to estimate density of F_. tritici could not be calculated using the Ruesink (1980) eguation because b values were greater than 2. Because adults of each species did not exhibit the same dispersion patterns, combining counts of all species in the flowers to estimate total density of thrips would not be statistically reliable. Further, each species differs in economic importance. The binomial sampling program appears feasible to estimate individual densities of adult thrips species. However, the implementation of this sampling program in areas where two or more species of Frankliniella occur simultaneously probably will not be practical.

PAGE 109

99 Although several species of thrlps Inhabit tomato flowers in North Florida, results from these studies indicate that cosmetic fruit damage is caused by oviposition from F_. occidental is . The other species apparently are not economically important in causing cosmetic fruit damage. Results also revealed that the relationship between the number of £. occidentalis female thrips in tomato flowers and the amount of damage (scars) was very variable. This expected to hamper afforts to develop economic injury levels for £. occidentalis in tomatoes. The focus of integrated pest management programs is to estimate population densities of individual species and apply control tactics only if economic thresholds are reached. The results of these studies provide a good example of the need of determining what species are present and their densities. The presence of several species in tomato fields and their differential economical importance needs that each species be managed individually. Additionally, several ecological characteristics such as seasonal distribution, spatial dispersion patterns, life cycles, etc. are different for each species. For the North Florida area £. occidentalis seems to be the key pest for both the cosmetic fruit damage and the transmission of TSWV. This species was the most abundant during the spring cropping season which was the season when the problem of thrips was more evident and caused considerable

PAGE 110

100 economic losses. Consequently, control tactics should be oriented to manage this species in the North Florida area. F. tritici was the second most abundant species during the spring cropping season and the most abundant one during the fall cropping season. However, this species does not transmit TSWV and does not cause cosmetic fruit damage. Consequently, F. tritici can not be considered as a pest of economic importance in tomatoes. Unfortunately, its similitude with F. occidental is and its relatively high populations in tomato flowers (37%) in North Florida indicates that it will cause confusion when monitoring for F. occidental is . Proper identification of thrips can not be done in the field. Thrips have to be carried to the laboratory and mounted on slides in order to identified them with a microscope. F_. fusca was present in low numbers only during the spring cropping season. This species is a vector of TSWV but is not important in causing cosmetic fruit damage. F. fusca can be easily separated from the other Frankliniella spp. even (under field conditions) , because of its dark brown color. Its economic importance will depend upon the presence of TSWV in the area. Because several species of thrips commonly inhabit tomato flowers in North Florida, the use of a binomial sampling program will have little value for integrated pest management programs. However, the binomial sampling programs will be

PAGE 111

101 useful for scouting programs in geographical areas where only one species commonly inhabit tomato flowers.

PAGE 112

REFERENCES Allen, W. R. and A. B. Broadbent. 1986. Transmission of tomato spotted wilt virus in Ontario greenhouses by Frankliniella occidental is . Can. J. Plant Pathol. 8(1) :33-38. Amin, P. W. , D. V. R. Reddy, and A. M. Ghanekar. 1981. Transmission of tomato spotted wilt virus, the causal agent of bud necrosis of peanut, by Scirtothrips dorsalis and Frankliniella schultzei . Plant Dis. 65 (8) : 663-665 . Ananthakrishnan, T. N. 1979. Biosystematics of thysanoptera . Ann. Rev. Entomol. 24:159-183. Ananthakrishnan, T. N. 1980. Thrips. In: Vector of plant pathogens. Edited by K.F. Harris and K. Maramorosch. Academic Press, N.Y. pp 149-164. Anonymous. 1985. Integrated pest management for tomatoes. Univ. Calif. Publication 3274. Second ed. 104 pp. Anonymous. 1989. Florida tomato committee, annual report 19881989. Orlando, Florida. 40 pp. Bailey, S. F. 1940. The distribution of injurious thrips in the United States. J. Econ. Entomol. 33 (1) : 133-136. Bailey, S. F. 1957. The thrips of California, Part I: Suborder Terebrantia. Bull. Calif. Ins. Surv. 4 (5): 143-220. Baker, J. R. and D. L. Stephan. 1986. Biology and control of the Western Flower Thrips. N.C. Flower Growers Bull. 30(4) :8-ll. Beshear, R. J. 1983. New records of thrips in Georgia (Thysanoptera, Terebrantia, Tubulifera) . J. GA. Entomol. Soc. 18(3) :342-344. Borror, D. J., D. M. Delong, and C. A. Triplehorn. 1976. An introduction to the study of insects. Holt, Rinehart, and Winstron. New York. Fourth edition. 852 pp. Bos, L. 1983. Introduction to plant virology. Longman, London 160 pp. 102

PAGE 113

103 Bottrell, D. R. 1979. Integrated pest management. Council on environmental quality. Superintendent of documents, U.S. Government Printing Office, Washington, D.C. 120pp. Boykin, L. S., W. V. Campbell, and M. K. Beute. 1984. Effect of pesticides on Neozvaites f loridana (Entomophthorales:Entomophthoraceae) and arthropd predators attacking the twospotted spider mite (AcarirTetranychidae) in North Carolina peanut fields. J. Econ. Entomol. 77 (4) : 969-975. Bryan, D. E. and R. F. Smith. 1956. The Frankliniella occidentalis (Pergande) complex in California (Thysanoptera: Thripidae) . Univ. Calif. Pub 1 . Entomol . 10(6) :359-410. Bullock, J. A. 1963. Extraction of Thysanoptera from samples of foliage. J. Econ. Entomol. 56 (5) : 612-614 . Cantelo, W. W. and R. E. Webb. 1980. Insect and diseases of vegetables in the home garden. USDA, Agric. inform. Bull. N.380. 54 pp. Carlson, E. C. 1964a. Effect of flower thrips on onion seed plants and a study of their control. J. Econ. Entomol. 57(5) :735-741. Carlson, E. C. 1964b. Damage to safflower plants by thrips and lygus and a study of their control. J. Econ. Entomol. 57(1) :140-145. Carlson, E. C. 1966. Further studies of damage to safflower plants by thrips and lygus bugs. J. Econ. Entomol. 59(1) :138-141. Carlson, E. C. and R. L. Witt. 1977. Insecticides for Frankliniella occidentalis and Lvaus hesperus on safflower plants. J. Econ. Entomol. 70 (4) : 460-462 . earner, G. R. , M. Shepard, and S. G. Turnipseed. 1974. Seasonal abundance of insect pests of soybeans. J. Econ. Entomol. 67(4) :487-493. Cho, J. J., R. f. L. Mau, T. L. German, R.W. Hartmann, L. S. Yudin, D. Gonzalves, and R. Prowidenti. 1989. A multidisciplinary approach to management of tomato spotted wilt virus in Hawaii. Plant Dis. 73 (5) : 375-383 . Cho, J. J., R. f. L. Mau, R. T. Hamasaki, and D. Gonzalves. 1988. Detection of tomato spotted wilt virus in individual thrips by enzyme-linked immunosorbent assay. Phytopathology. 78 (10) : 1348-1352 .

PAGE 114

104 Cho, J. J., w. C. Mitchell, and R. Mau. 1986. Development of control procedures for tomato spotted wilt virus (TSWV) disease. Phytopathology. 76(10) :1134. Cho, J. J., W. C. Mitchell, L. Yudin, and L. Takayama. 1984. Ecology and epidemiology of tomato spotted wilt virus (TSWV) and its vector Frankliniella occidental is . Phytopathology. 74(7) :866. Clough, G. H. 1987. Florida tomato industry. Florida Agric. Inform. Retrieval Syst. (FAIRS) . Tomato Production Database. IFAS. University of Florida, Gainesville. Dintenfass, L. P., D. P. Bartell, and M. A. Scott. 1987. Predicting resurgence of Western flower thrips (Thysanoptera:Thripidae) on onions after insecticide application in the Texas high plains. J. Econ. Entomol. 80(2) :502-506. Duncan, D. B. 1955. Multiple range and multiple F tests. Biometrics. 11:1-42. Eddy, C. 0. and E. M. Livingstone. 1931. Frankliniella fusca Hinds (thrips) on seedling cotton. S. C. Agric. Exp. Stn. Bull. 271:1-23. Elliot, N. C. and R. W. Kieckhefer. 1987. Spatial distribution of cereal aphids (Homoptera:Aphididae) in winter wheat and spring oats in South Dakota. Environ. Entomol. 16(4) :896-901. Funderburk, J. 1988. Thrips in spring tomatoes. Panhandle Agri-News. IFAS. University of Florida. 10(2) :2-4. Funderburk, J. E. and T. P. Mack. 1987. Abundance and dispersion of Geocoris spp. (Hemiptera: Lygaeidae) in Alabama and Florida soybean fields. Fla. Entomol. 70(4) :432-439. Funderburk, J. E. and T. P. Mack. 1989. Seasonal abundance and dispersion patterns of damsel bugs (Hemiptera :Nabidae) in Alabama and Florida soybean fields. J. Entomol. Sci. 24(1) :9-15. Gerrard, D. J. and H. C. Chiang. 1970. Density estimation of com rootworm egg populations based upon freguency of occurrence. Ecology. 51:238-245. Gerrard, D. J . and R. D. Cook. 1972. Inverse binomial sampling as a basis for estimating negative binomial population densities. Biometrics. 28:971-980.

PAGE 115

105 Gonzalez, D. , B. R. Patterson, T. F. Leigh, and L. T. Wilson. 1982. Mites: a primary food source for two predators in San Joaquin Valley cotton. Calif. Agric. 36:18-20 Graves, J. B. , J. D. Powell, M. E. Farris, S. Micinski and R. N. Story. 1987. Western flower thrips, a new cotton pest in Louisiana, Louisiana Agric. 30(4) :4-8. Greenough, D. R. , L. L. Black, R. N. Story, and L. D. Newsom. 1985. Occurrence of Frankliniella occidental is in Louisiana: a possible cause for the increased incidence of tomato spotted wilt virus. Phytopathology. 75(11) :1362. Hagan, A. K. , J. R. Weeks, and J. C. French. 1987. Identification and control of tomato spotted wilt on peanut. Agriculture and Natural Resources Information, Ala. Coop. Ext. Serv. , Auburn Univ. Harding, J. A. 1961. Effect of migration, temperature, and precipitation on thrips infestations in south Texas. J. Econ. Entomol. 54(1): 77-79. Henneberry, T. J., F .F. Smith, and D. Shriver. 1964. Flower thrips in outdoor rose fields and an improved method of extracting thrips from rose flowers. J. Econ. Entomol. 57(3) :410-412. Hutchins, S. H., L. G. Higley, and L. P. Pedigo. 1988. Injury equivalency as a basis for developing multiple-species economic injury levels. J. Econ. Entomol. 81(1) :l-8. Ie, T. S. 1970. Tomato spotted wilt virus. N. 39. Description of plant viruses. Commonw. Mycol. Inst. /Assoc. Appl. Biol., Kew, Surrey, England. 3pp. Irwin, M. E. and K. V. Yeargan. 1980. Sampling phytophagous thrips on soybean. In: Sampling methods in soybean entomology, Edited by M. Kogan and D. C. Herzog. Springer-verlag, New York. pp. 284-304. Irwin, M. E., K. V. Yeargan, and N. L. Marston. 1979. Spatial and seasonal patterns of phytophagous thrips in soybean fields with comments on sampling techniques. Environ. Entomol. 8(1) :131-140. Jensen, F. 1973. Flower thrips damage to table grapes in San Joaquin Valley: (1) Timing of halo spotting by flower thrips on table grapes. Calif. Agric. 27:6-8.

PAGE 116

106 Jensen, F. and D. Luvisi. 1973. Flower thrips damage to table grapes in San Joaquin Valley: (2) Flower thrips nymphs involved in scarring of Thompson seedless grapes. Calif. Agric. 27:8-9. Jones, R. K. and J. W. Moyer. 1986. Tomato spotted wilt virus in Gloxinia in North Carolina. N. C. Flower Growers Bull. 30(4) :11-13. Karadinos, M. G. 1976. Optimum sample size and comments on some published formulae. Bull. Entomol. Soc. Amer. 22:417-421. Karny, H. 1910. Neve thysanopteren der wiener gegend. Naturw. Ver. Wien. Mitt. 8:41-57. Kinzer, R. E. 1968. Mass rearing the tobacco thrips, Frankliniella fusca (Hinds) , and laboratory techniques for testing peanut resistance to thrips. Master's thesis Oklahoma State University. 13 pp. Kinzer, R. E. , S. Young, and R. R. Walton. 1972. Reading and testing tobacco thrips in the laboratory to discover resistance in peanuts. J. Econ. Entomol. 65 (3) : 782-785. Kono, T., and C. S. Papp. 1977. Handbook of agricultural pests, aphids, thrips, mites, snails, and slugs. Depart. Food and Agric. Div. Plant Ind. St. Calif. 205 pp. Kucharek, T. 1986. Tomato spotted wilt virus found in Florida in 1986. Proc. Fla. Tomato Inst. Veg. Crops Ext. Rep. VEC86-1: 50-51. Lange, W. H. and L. Bronson. 1981. Insect pests of tomatoes. Ann. Rev. Entomol. 26:345-371. Laster, M. L. , R. S. Baker, and W. F. Kitten. 1984. Effects of dinoseb and dinoseb + MSMA on arthropod populations in cotton fields. J. Econ. Entomol. 77 (3) :741-743. Letourneau, D. K. and M. A. Altieri. 1983. Abundance patterns of a predator, Orious tristicolor (Hemiptera: Anthocoridae) , and its prey, Frankliniella occidental is (Thysanoptera: Thripidae) : Habitat attraction in polycultures versus monocultures. Environ. Entomol. 12(5) :1464-1469. Lewis, T. 1969a. The diversity of the insect fauna in a hedge-row and neighbouring fields. J. Appl. Ecol. 6:453-458.

PAGE 117

107 Lewis, T. 1969b. The distribution of flying insects near a low hedge row. J. Appl. Ecol. 6:443-452. Lewis, T. 1973. Thrips their biology, ecology and economic importance. Academic Press, London. 349 pp. Lublinkhof, J. and D. E. Foster. 1977. Development and reproductive capacity of Frankliniella occidental is (Thysanoptera:Thripidae) reared at three temperatures. J. Kansas Entomol. Soc. 50:313-316. Lynch, R. E., J. W. Garner, and L. W. Morgan. 1984. Influence of systematic insecticides on thrips damage and yield of Florunner peanuts in Georgia. J. Agric. Entomol. 1:33-42. Maiteki, G. A., R. J. Lamb, and S. T. Ali-khan. 1986. Seasonal abundance of the pea aphid, Acyrthosiphon pi sum (Homoptera: Aphididae) , in manitoba field peas. Can. Entomol. 118:601-607. Marston, N. L. , G. D. Thomas, C. M. Ignoffo, M. R. Gebhardt, D. L. Hostetter, and W. A. Dickerson. 1979. Seasonal cycles of soybean arthropods in Missouri: Effect of pesticidal and cultural practices. Environ. Entomol. 8(1) :165-173. Matthews, R. E. F. 1970. Plant virology. Academic Press, New York. 778 pp. McRitchie, J. J. 1986. Tomato spotted wilt. Plant pathology circular No. 287. Division of Plant Industry, Florida. 2 pp. Morgan, A. C. 1913. New genera and species of Thysanoptera with notes on distribution and food plants. Proc. U. S. Nat. Museum. 2008 (46) : 1-55. Morgan, L. W. , J. w. Snow, and M. J. Peach. 1970. Chemical thrips control: effects on growth and yield of peanuts in Georgia. J. Econ. Entomol. 63 (4) : 1253-1255. Newsom, L. D., J. S. Roussel, and C. E. Smith. 1953. The tobbaco thrips, its seasonal history and status as a cotton pest. Louisiana Agric. Exp. Sta. Tech. Bull. 474:36 pp. Nugaliyadele, L. and E. A. Heinrichs. 1984. Biology of rice thrips, Stenchaetothrips biformic (Bagnall) (Thysanoptera :Thripidae) and a greenhouse rearing technique. J. Econ. Entomol. 77 (5) : 1171-1175.

PAGE 118

108 Nyrop, J. P., A. M. Agnello, J. Kovach, and W. H. Reissig. 1989. Binomial sequential classification sampling plans for european red mite (Acari: Tetranychidae) with special reference to performance criteria. J. Econ. Entomol. 82(2) :482-490. Oetting, R. D. 1985. Insecticide efficacy against western flower thrips, Georgia, 1985. Insect ic. and Acaricide Tests. 11:196. Olson, S. M. and J. E. Funderburk. 1986. New threatening pest in Florida western flower thrips. Proc. Fla. Tomato Inst. Veg. Crops Ext. Rep. VEC 86-1. Ota, A. K. 1968. Comparison of three methods of extracting the flower thrips from rose flowers. J. Econ. Entomol. 61(6) :1754-1755. Paliwal, Y. C. 1976. Some characteristics of the thrips vector relationship of tomato spotted wilt virus in Canada. Can. J. Bot. 54 (5) :402-405. Paliwal, Y. C. 1979. Occurrence and localization of spherical virus like particles in tissues of apparently healthy tobacco thrips Frankliniella fusca, a vector of tomato spotted wilt virus. J. Invertebr. Pathol. 33:307-315. Palmer, J. M. , L. A. Mound, and G. J. du Heaume. 1989. 2. Thysanoptera . In: CIE Guides to insects of importance to man. Edited by C. R. Betts. C.A.B. Internat. Inst. Entomol. Brit. Mus. Nat. Hist. 75 pp. Pedigo, L. P. 1989. Entomology and pest management. MacMillan Publishing Company, New York. 646 pp. Pedigo, L. P., S. H. Hutchins, and L. G. Higley. 1986. Economic injury levels in theory and practice. Ann. Rev. Entomol. 31:341-368. Pena, J. E. 1983. Tomato pinworm, Keiferia lvcopersicella (Walsingham) : population dynamics and assessment of plant injury in southern Florida. Ph.D. dissertation, University of Florida. 265 pp. Pickett, C. H., L. T. Wilson, and D. Gonzalez, 1988. Population dynamics and within-plant distribution of the western flower thrips (Thysanoptera : Thripidae) , an early-season predator of spider mites infecting cotton. Environ. Entomol. 17 (3) : 551-559 .

PAGE 119

109 Pohronezny, K. , V. H. Waddill, D. J. Schuster, and R. M. Sonoda. 1986. Integrated pest management for Florida tomatoes. Plant Dis. 70 (2) : 96-102 . Price, J. E. , L. D. Ketzler, and C. D. Stanley. 1981. Sampling methods for Liriomyza trifolii . In: Proceedings of IFASindustry conference on biology and control of Liriomyza leafminers. Ed. by D. J. Schuster. University of Florida, Gainesville, pp 141-149. Race, S. R. 1961. Early season thrips control on cotton in New Mexico. J. Econ. Entomol. 54 (5) :974-976. Race, S. R. 1965. Predicting thrips populations on seedling cotton. J. Econ. Entomol. 58 (5) : 1013-1014 . Reddick, B. B. , C. H. Hadden, S. C. Bost, and M. A. Newman. 1987. First report of tomato spotted virus in Tenessee. Plant Dis. 71:376. Robb, K. L. , J. P. Newman, and M. P. Parrella. 1987. The biology and control of the Western flower thrips. Compilation of two presentations given at the 1987 OSFA short course detailing the biology and control of WFT. Depart. Entomol. Univ. Calif. Riverside, Ca. 19 pp. (unpublished) . Ruesink, W. G. 1980. Introduction to sampling theory. In: Sampling methods in soybean entomology. Ed. by M. Kogan and D. C. Herzog. Springer-Verlag, New York. pp. 61-78. Ruesink, W. G. and M. Kogan. 1975. The quantitative basis of pest management: sampling and measuring. In: Introduction to insect pest management. Ed. by R. L. Metcalf and W. H. Luckmann. pp. 309-351. ' Rummel, D. R. and J. E. Quisenberry. 1979. Influence of thrips injury on leaf development and yield of various cotton genotypes. J. Econ. Entomol. 72 (5) :706-709. Sakimura, K. 1953. Frankliniella tritici, a non-vector of the spotted wilt virus. J. Econ. Entomol. 46 (5) : 915-916. Sakimura, K. 1962. Frankliniella occidentalis (Thysanoptera: Thripidae) , a vector of the tomato spotted wilt virus, with special reference to the color forms. Ann. Entomol. Soc. Amer. 55:387-389. Sakimura, K. 1963. Frankliniella f usca . an additional vector for the tomato spotted wilt virus with notes on Thrips tabaci, another vector. Phytopathology. 53:412-415.

PAGE 120

110 Sakimura, K. and K. O'Neill. 1979. Frankliniella. redefinition of genus and revision of minuta group species (ThysanopterarThripidae) . Tech. Bull. N.1572. USDA. 49 pp. Salguero, V. and J. Funderburk. 1989. Flower thrips in spring and fall tomatoes. Pandhandle Agri-News. IFAS University of Florida. 10(2) :2-4. Schuster, D. J. and J. F. Price. 1987. The western flower thrips and the sweetpotato whitefly: New pests threatening Florida tomato production. Bradenton Res. Rep. BRA 1987-16. 7 pp. Shepard, M. , G. R. earner, and S. G. Tumipseed. 1974. Seasonal abundance of predaceous arthropods in soybeans. Environ. Entomol. 3 (6) : 985-988 . Smith, D. H. 1979. Disease resistance in peanuts. In: Biology and breeding for resistance to arthropods and pathogens in agricultural plants. Ed. by M. K. Harris. Proceedings of a short course in host plant resistance. Texas A&M University, pp. 431-447. Southwood, T. R. E. 1978. Ecological methods. Chapman & Hall. London. 524 pp. Sprenkel, R. K. 1987. Western flower thrips crops infested and control. Triennial extension entomology plant pathology waorkshop for the southern region. Oct. 4-7, 1987. Hyatt Regency, Savannah, Georgia. 9 pp. (unpublished) . Sprenkel, R. K. 1988. TSWV and thrips on tomatoes. IPM Newsletter. IFAS. University of Florida. 4:3. Sprenkel, R. K. 1989. TSWV found on peanuts. IPM Newsletter. IFAS. University of Florida. 7:2. Stafford, E. M. 1974. Insect grape pests of northern California. Calif. Agric. Exp. St. circular 566. 75 pp. Stannard, L. J. 1968. The thrips, or Thysanoptera of Illinois. 111. Nat. Hist. Surv. Bull. 29(4) :552 pp. Steel, R. G. D. and J. H. Torrie. 1960. Principles and procedures of statistics. McGraw-Hill Book Company, New York. 481 pp. Stern, V. M. , R. F. Smith, R. Van den Bosch, and K. S. Hagen. 1959. The integrated control concept. Hilgardia. 29:81101.

PAGE 121

Ill Stewart, J. W. , C. Cole, and P. Lummus. 1989. Winter survey of thrips (Thysanoptera:Thripidae) from certain suspected and confirmed hosts of tomato spotted wilt virus in south Texas. J. Entomol. Sci. 24 (3) : 392-401. Stoltz, R. L, and V. M. Stern. 1978a. Cotton arthropod food chain disruption by pesticides in the San Joaquin Valley, California. Environ. Entomol. 7 (5) :703-707. Stoltz, R. L. and V. M. Stem. 1978b. The longevity and fecundity of Orius tristicolor when introduced to increasing numbers of their prey Frankliniella occidental is . Environ. Entomol. 7(2) :197-198. Tappan, W. B. 1986a. Relationship of sampling time to tobacco thrips (Thysanoptera: Thripidae) numbers in peanut foliage buds and flowers. J. Econ. Entomol. 79(5) :1359-1363. Tappan, W. B. 1986b. Tobacco thrips (Thysanoptera: Thripidae) number after peanut foliage bud and flower excision. J. Econ. Entomol. 79 (4) : 1082-1084 . Tappan, W. B. and D. W. Gorbet. 1979. Relationship of seasonal thrips populations to economics of control on Florunner peanuts in Florida. J. Econ. Entomol. 72 (5) :772-776. Tappan, W. B. and D. W. Gorbet. 1981. Economics of tobacco thrips control with systemic pesticides on Florunner peanuts in Florida. J. Econ. Entomol. 74 (3) : 283-286. Taylor, L. R. , I. P. Woiwood, and J. N. Perry. 1978. The density-dependence of spatial behavior and the varity of randomness. J. Anim. Ecol. 47:383-406. Taylor, R. L. 1961. Aggregation, variance, and the mean. Nature. 189: 732-735. Terry, L. I. and G. DeGrandi-Hof fman. 1988. Monitoring western flower thrips (Thysanoptera: Thripidae) in "Granny Smith" apple blossom clusters. Can. Entomol. 120:1003-1016. Tingey, W. M. 1979. Breeding for arthropod resistance in vegetables. In: Biology and breeding for resistance to arthropods and pathogens in agricultural plants. Ed. by M. K. Harris. Proceedings of an international short course in host plant resistance. Texas A&M University, pp. 495-522.

PAGE 122

112 Trichilo, P. J. and T. F. Leigh. 1986. Predation on spider mite eggs by the western flower thrips, Frankliniella occidental is (Thysanoptera:Thripidae) , an oportunist in a cotton agroecosystem. Environ. Entomol. 15 (4) : 821-825. Walkey, D. G. A. 1985. Applied plant virology. John Wiley & sons, New York. 329 pp. Waters, W. E. 1959. A quantitative measure of aggregation in insects. J. Econ. Entomol. 52 (6) : 1180-1184. Watson, J. R. 1922. The proper name and distribution of the Florida flower thrips. Fla. Entomol. 7(1):9-11. Watson, J. R. 1923. Synopsis and catalog of the Thysanoptera of North America. Univ. Fla. Agric. Expt. Sta. Bull., 168. 100 pp. Watson, T. F. 1965. Influence of thrips on cotton yields in Alabama. J. Econ. Entomol. 58 (6) : 1118-1122 . Watts, J. G. 1934. A comparison of the life cycles of Frankliniella tritici (Fitch) , F_. fusca (Hinds) , and Thrips tabaci Lind. (Thysanoptera: Thripidae) in South Carolina. J. Econ. Entomol. 27(6) :1158-1159. Watts, J. G. 1936. A study of the biology of the flower thrips Frankliniella tritici (Fitch) with special reference to cotton. S. C. Agric. Exp. Stn. Bull. 306:1-46. Watts, J. G. 1937a. Reduction of cotton yields by thrips. J. Econ. Entomol. 30 (6) : 860-863 . Watts, J. G. 1937b. Species of thrips found on cotton in South Carolina. J. Econ. Entomol. 30(6) :857-860. Webb, R. E., S. W. Jacklin, G. V. Johnson, J. W. Mackley, and E. J. Paugh. 1970. Seasonal variation in population of flower thrips in Georgia, Meryland, and New York. J. Econ. Entomol. 63 (5) : 1392-1394 . Wilson, L. F. and D. J. Gerrard. 1971. A new procedure for rapidly estimating european pine sawfly (Hymenoptera: Diprionidae) population levels in young pine plantations. Can. Entomol. 103 (9) : 1315-1322 . Wilson, L. T. 1982. Growth and fruiting development of normal and terminal damaged cotton plants. Environ. Entomol. 11(2) :301-305.

PAGE 123

113 Wilson, L. T. 1985. Estimating the abundance and impact of arthropod natural enemies in IPM systems. In: Biological control in agricultural IPM systems. Ed. by M. A. Hoy and D. C. Herzog. Academic Press, Inc. Orlando, Florida, pp. 303-322. Wilson, L. T., T. F. Leigh, and V. Maggi. 1981. Presenceabsence sampling of spider mite densities on cotton. Calif. Agric. 35:10. Wilson, L. T., C. Pickel, R. C. Mount, and F. G. Zalom. 1983. Presence-absence sequential sampling for cabbage aphid and green peach aphid (Homoptera: Aphididae) on brussels sprouts. J. Econ. Entomol. 76(3) : 476-479. Wilson, L. T. and P. M. Room. 1983. Clumping pattern of fruit and arthropods in cotton, with implications for binomial sampling. Environ. Entomol. 12(l):50-54. Yokoyama, V. Y . 1977. Frankliniella occidentalis and scars on table grapes. Environ. Entomol. 6(l):25-30. Young, O. P. and W. C. Welbourn. 1988. Parasitism of Trigonotylus dady (Heteroptera: Miridae) by Las ivervthraeus iohnstoni (Acari: Esrythracidae) , with notes on additional hosts and distribution. J. Entomol. Sci. 23 (3) :269-273. Yudin, L. S., J. J. Cho,and W. C. Mitchell. 1986. Host range of western flower thrips, Frankliniella occidentalis (Thysanoptera: Thripidae) with special reference to Leucaena glauca . Environ. Entomol. 15(6):1292-5 Yudin, L. S., W. C. Mitchell, and J. J. Cho. 1987. Color preference of thrips (Thysanoptera: Thripidae) with reference to aphids (Homoptera: Apididae) and leaf miners in Hawaiian lettuce farms. J. Econ. Entomol. 80(l):51-55. Yudin, L. S., B. E. Tabashnik, J. J. Cho, and W. C. Mitchell. 1988. Colonization of weeds and lettuce by thrips (Thysanoptera : Thripidae) . Environ. Entomol. 17(3):522-6. Zalom, F. G., M. A. Hoy, L. T. Wilson, and W. W. Barnet. 1984. Presence-absence sequential sampling for Tetranychus mite species. Hilgardia. 52(7):14-24. Zur Strassen, V. R. 1986. Frankliniella occidentalis (Pergande 1985) from North America as a new thysanopterous inhabitant of European greenhouses. Nachrichtenbl . Deut. Pflanzenschutzd. , 38(6):86-88.

PAGE 124

BIOGRAPHICAL SKETCH Victor Eberto Salguero Navas was born in El Progreso, Jutiapa, Guatemala in 1952. He received a high school degree as professor of elementary schools in rural areas in 1970. Upon graduation, he attended the University of San Carlos de Guatemala and received the Bachelor of Science degree in agronomy (Ingeniero Agronomo) in 1977. Since 1976 he has been working at the Instituto de Ciencia y Tecnologia Agricolas (ICTA) in Guatemala, where he has carried out agronomic research activities focused mainly in the development of resistant varieties of beans to Bean Golden Mosaic Virus (BGMV) and to the bean weevil Apion godmani . Simultaneously he was a part-time professor of Entomology at the University Rafael Landivar in Jutiapa, Guatemala . He received his master's degree in crop protection from New Mexico State University in 1981 under the direction of Dr. Joe Ellington. He enrolled in the Department of Entomology and Nematology at the University of Florida in 1986, pursuing the Ph.D. degree under the direction of Dr. Joe Funderburk. His research is being conducted in north Florida on the biology and management of flower thrips, Frankliniella spp., in tomato fields. 114

PAGE 125

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the decree of Doctor of Philosophy. Donald C. Herzoq, Professor of Entomo atology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Joseph E. Funderburk, Cochairman Assistant Professor of Entomology and Nematology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and/c?uality , a a dissertation for the degree pty Doctor of Philosophy. Richard K. Sprenkel Associate Professo, Nematology Entomology and I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. R. Strayer essor of Entomology and Nematology I certify that I have read this study and that in my opinion in conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Steven M. Olson Associate Professor of Horticultural Science

PAGE 126

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May 1990 Dean, College of Agriculture jtacj X. thy ,, Crallege of Aq£j/c\ Dean, Graduate School