Temperature-dependent development and influence of larval instars of Liriomyza sativae Blanchard on parasitization by Op...

MISSING IMAGE

Material Information

Title:
Temperature-dependent development and influence of larval instars of Liriomyza sativae Blanchard on parasitization by Opius dissitus Muesebeck
Physical Description:
ix, 137 leaves : ill. ; 28 cm.
Language:
English
Creator:
Petitt, Frederick Lewis, 1956
Publication Date:

Subjects

Subjects / Keywords:
Opius   ( lcsh )
Liriomyza   ( lcsh )
Pests -- Biological control   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1988.
Bibliography:
Includes bibliographical references (leaves 99-112).
Statement of Responsibility:
by Frederick Lewis Petitt.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001475661
notis - AGY7482
oclc - 20867877
sobekcm - AA00004823_00001
System ID:
AA00004823:00001

Full Text











TEMPERATURE-DEPENDENT DEVELOPMENT AND INFLUENCE OF
LARVAL INSTARS OF Liriomyza sativae BLANCHARD ON
PARASITIZATION BY Opius dissitus MUESEBECK










BY

FREDERICK LEWIS PETITT


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


1988














ACKNOWLEDGEMENTS


I would like to gratefully acknowledge all members of

my supervisory committee for their guidance, advice, and

encouragement: Drs. Carl Barfield, Pauline Lawrence, Jon

Allen, Gary Leibee, and Jerry Stimac. The assistance of Dr.

Tom Ashley is also appreciated.

I would also like to thank the Walt Disney Company for

financial support and the staff at The Land, EPCOT Center

for their help. Dr. Hank Robitaille and the senior staff

provided support and encouragement. Eldon Muller designed

the system used to collect and process environmental data

and Andrew Schuerger provided valuable discussions. The

entomological staff at The Land over the years, Marian Cof-

fey, Chris Halliday, Debbie Karan, Maureen Knop, Bernadette

Scott, and David Wietlisbach, all helped in various ways.

Dinah Jordan typed the manuscript.

I would like to acknowledge Dr. Ken Spencer and Dr. Bob

Wharton for identifying specimens of L. sativae and 0.

dissitus, respectively, and Dr. Frank Martin for his assis-

tance with statistical analyses.

I also would like to express my sincere appreciation to

my family for encouragement, patience, and sacrifice.















TABLE OF CONTENTS

Pacge
ACKNOWLEDGEMENTS......................................................ii

LIST OF TABLES............................... ....... ...... v

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

ABSTRACT................... ........................... viii

CHAPTER

I. INTRODUCTION.............................. 1

Taxonomy of L. sativae and L. trifolii.... 2
Host Range of L. sativae................... 3
Biology and Life Cycle of L. sativae...... 4
Integrated Pest Management for Liriomyza
Leafminers.. .............................. 5
Biology of Parasitoids of Liriomvza spp... 5
Biological Control of Liriomyza spp. in
Greenhouses................................ 6
Rearing Opius dissitus..................... 7

II. DISTINGUISHING LARVAL INSTARS OF Liriomvza
sativae.............. ............ ....... 10

Introduction .............................. 10
Materials and Methods....................... 12
Results and Discussion..................... 14
Conclusions..... ....................... ... 22

III. EFFECTS OF TEMPERATURE ON DEVELOPMENT OF
IMMATURE STAGES OF Liriomvza sativae...... 24

Introduction............................... 24
Materials and Methods..................... 26
Results............. ..................... .. 30
Discussion.... ............................ 38
Conclusions................................ 42


iii










IV. INTRASPECIFIC COMPETITION AMONG Liriomyza
sativae LARVAE IN PRIMARY LEAVES OF LIMA
BEAN.................. ................... 44

Introduction.............................. 44
Materials and Methods........................ 45
Results............. ............. ...o ....... 48
Discussion. *...................... ....... 53
Conclusions................................ 55

V. EFFECTS OF PARASITIZATION OF Liriomvza
sativae INSTARS ON NUMBER AND CHARACTER-
ISTICS OF OQius dissitus PROGENY.......... 57

Introduction............................. 57
Materials and Methods...................... 58
Results.................... ....... ... ....... 63
Discussion............................... .. 67
Conclusions................................. 71

VI. EFFECTS OF TEMPERATURE AND HOST INSTAR ON
DEVELOPMENT TIME OF Opius dissitus......... 73

Introduction.................................. 73
Materials and Methods..................... 74
Results..... ........ ........... .. .... 78
Discussion... .............................. o.88
Conclusions....................................... 92

VII. CONCLUSIONS................................ 93

REFERENCES CITED............................... ...... 99

APPENDICES

A. MODIFIED HOAGLAND NUTRIENT SOLUTION... 114

B. DISTRIBUTED DELAY PROGRAM.............. 115

C. SUMMARIZED DATA ON DEVELOPMENT OF
IMMATURE STAGES OF Liriomvza
sativae................................. 130

D. INSTAR DETERMINATIONS FOR LARVAE USED
IN EXPERIMENTS IN CHAPTER V......... 133

E. INSTAR DETERMINATIONS FOR LARVAE USED
IN EXPERIMENTS IN CHAPTER VI.......... 136


BIOGRAPHICAL SKETCH............................. ....... 137















LIST OF TABLES


Table Page

2.1. Length of mouth hooks and cephalopharyngeal
skeleton for larval instars of L. sativae....... 16

2.2. Growth ratios of the length of cephalopharyn-
geal skeleton of L. sativae as compared with
larvae of other Liriomyza species.............. 20

2.3. Growth ratios of the length of mouth hooks of
L. sativae as compared with other Liriomyza
species.......................................... 21

3.1. Linear regression equations for relationships
between development rate and temperature for
the egg, larval and pupal stages of L. sativae
on lima bean primary leaves.................... 32

3.2. Regression equations for development rate and
degree-days (DD) required from oviposition to
the beginning of each instar and to larval
emergence........................................ 36

3.3. Percentage of the total egg-larval development
time spent in each stage or instar for L.
sativae, L. brassicae and L. trifolii........... 41

4.1. Influence of L. sativae larval density on mor-
tality in primary leaves of lima bean............ 49

4.2. Influence of L. sativae larval density on
weight and percentage survival of pupae......... 51

4.3. Comparison of pupal weights for L. sativae lar-
vae reared at high and low densities............. 52

5.1. Numbers of 0. dissitus progeny produced when
parental females exposed to near equal densi-
ties of different host instars.................. 65

5.2. Size of 0. dissitus progeny resulting from par-
asitization of various larval instars of L.
sativae......................................... 66








5.3. Mean number of chorionated eggs in female
progeny of 0. dissitus resulting from parasiti-
zation of different instars of L. sativae....... 68

6.1. Egg-to-adult development times for 0. dissitus
resulting from parasitization of different
instars of L. sativae at 250C................ 80

6.2. Egg-to-adult development times for 0. dissitus
on a mixture of second and third instar L.
sativae at different constant temperatures...... 82

6.3. Time elapsed from 0. dissitus oviposition to
emergence of parasitized and unparasitized L.
sativae larvae from leaves....................... 86

6.4. Timing of adult eclosion for L. sativae and 0.
dissitus at 250C when a mixture of second and
third instar hosts were parasitized............. 89














LIST OF FIGURES


Figure Page

2.1. Relationship between lengths of mouth hooks and
cephalopharyngeal skeleton for L. sativae....... 15

2.2. Cephalopharyngeal skeleton including mouth
hooks of first, second, and third instar larvae
of L. sativae................................. ... 18

3.1. Relationship of development time and develop-
ment rate to temperature for the egg and larval
stages of L. sativae on bush lima bean........... 31

3.2. Relationship of development time and develop-
ment rate to temperature for the pupal stage of
L. sativae reared on bush lima bean............. 33

3.3. Relationship of development rates to tempera-
ture for larval instars of L. sativae reared on
bush lima bean.................................. 34

3.4. Comparison of model output and observed devel-
opment of L. sativae in fluctuating temperature
experiment................................... ...... 37

3.5. Larval development time for Liriomyza species
on various host plants......................... 40

6.1. Percentage of 0. dissitus in the first stadium
within L. sativae parasitized either as first
or third instar larvae............................. 81

6.2. Development time of L. sativae and 0. dissitus
in constant temperature experiments from 15 to
350C........ -....... .... ....... .............. 85

6.3. Frequency distribution for time of day of adult
eclosion for L. sativae and 0. dissitus.......... 87


vii














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

TEMPERATURE-DEPENDENT DEVELOPMENT AND INFLUENCE OF
LARVAL INSTARS OF Liriomyza sativae BLANCHARD ON
PARASITIZATION BY Opius dissitus MUESEBECK

Frederick Lewis Petitt

December 1988

Chairperson: Carl S. Barfield
Cochairperson: Pauline 0. Lawrence
Major Department: Entomology and Nematology

Experiments were conducted to develop methods for rear-

ing ODius dissitus Muesebeck, a parasitoid of the vegetable

leafminer, Liriomyza sativae Blanchard, on primary leaves of

the lima bean plant. Methods were developed to distinguish

larval instars of L. sativae by measurement of the cephalo-

pharyngeal skeleton. Development times were determined for

eggs and each larval instar at constant temperatures from 20

to 350C. These data were used to model development using a

distributed delay. The model accurately predicted timing of

egg eclosion and larval molts under fluctuating temperature

conditions.

The larval development model was employed to design

experiments in which equal densities of first, second, and

third instar L. sativae larvae were exposed to individual Q.

dissitus females. Exposure of second and third instar lar-

vae yielded 150 to 200% more progeny, respectively, than did


viii








first instar larvae. Consequently, second and third instar

larvae were used in all subsequent rearing efforts. Incuba-

tion periods required to reach these stadia were determined

for fluctuating temperature conditions using the model.

Competition between _L. sativae larvae was studied to

determine acceptable rearing densities. Exploitative compe-

tition among larvae in the same cohort was minimized by

using densities below 0.8 third instar larvae/cm2

0. dissitus developed more quickly than L. sativae at

temperatures of 200C or higher. Egg-to-adult development

rates (y) of the parasitoid on second and third instar hosts

at temperatures (x) from 15 to 300C were described by the

equation: y = 0.0052x 0.0558.

The parasitoid required 0.5-2d longer to develop on

second and first instar hosts, respectively, than on third

instar hosts at 250C. The delay on younger hosts was due to

prolongation of the first larval stadium of the parasitoid.

Regardless of the host instar parasitized, 73% or more of

the parasitoids underwent the first larval molt between 24

and 48h after host pupariation.













CHAPTER I
INTRODUCTION


Most members of the genus Liriomyza consume leaf meso-

phyll tissue during larval development. Several species

form serpentine mines and have been called serpentine leaf-

miners (Steyskal 1973). Damage to various ornamental and

vegetable crops by serpentine leafminers has increased dra-

matically during the last 10 years (Parrella & Keil 1984).

The importance of the problem is exemplified by the fact

that five conferences were held during the past seven years

specifically to address this problem (1980 (conference pro-

ceedings not published), Schuster 1981, Poe 1982, Poe 1984,

Knodel-Montz 1985). Unsuccessful leafminer control has been

due largely to rapid acquisition of resistance to insecti-

cides (Leibee 1981a). In addition, the frequent use of

broad-spectrum pesticides has reduced populations of native

parasitoids (Wene 1953, Oatman & Kennedy 1976, Johnson et

al. 1980a,b).

Although most Liriomyza species are restricted to a

single or small group of hosts, 12 species feed on a wide

range of hosts (Spencer & Steyskal 1986). Three species,

Liriomvza sativae Blanchard, L. trifolii (Burgess), and L.

huidobrensis (Blanchard) are the most serious pests in the

United States (Spencer & Steyskal 1986). While L.

huidobrensis is reported to be common only in southern

1







2
California, I. sativae and L. trifolii occur widely through-

out southern states from Florida to California (Spencer &

Steyskal 1986).


Taxonomy of L. sativae and L. trifolii

The taxonomic history of Liriomyza is riddled with con-

fusion and the taxonomy of L. sativae and L. trifolii is no

exception. Only the misidentifications and confusion

involving L. sativae will be discussed here. A taxonomic

history of all economically important Liriomyza in the

United States is available elsewhere (Spencer 1981b).

L. sativae was first described in 1938 occurring on

alfalfa (Medicago sativa L.) in Argentina (Spencer 1981a).

In 1957 Frick correctly described this species as L. munda,

and he also incorrectly applied the name Aqromyza pictella

(which had been described originally by Thomson in 1869) to

specimens later determined by Spencer to be L. sativae

(Spencer 1981b). The name L. pictella Thomson was used by

Oatman to refer to the "melon leaf miner," on which he and

Michelbacher conducted several biological and ecological

studies (Oatman & Michelbacher 1958, 1959, Oatman 1959a,b,

1960). Spencer (1981b) suspects that Oatman's "melon leaf

miner" was L. sativae. Oatman (1961) carried out cross-

breeding studies between the "melon leaf miner" and the

"tomato leaf miner" (L. munda Frick) and reported that "no

structural differences were found either in the male geni-

talia or larval spiracles" and that each of the two cultures

would transfer to the host of the other.








L. sativae and L. trifolii can be confused easily

because of their similar morphology and overlapping host

ranges. Spencer (1965) separated these species (1. sativae

was then referred to as L. munda) on the basis of whether

the vertical bristles (between ocellar triangle and eye) are

on dark ground (L. sativae) or yellow ground (L. trifolii),

and whether the mesonotrum is a brilliant, shining black (L.

sativae) or blackish-gray and pollinose (L. trifolii).


Host Range of L. sativae

A list of confirmed hosts of L. sativae in Florida has

been reported (Stegmaier 1966). The host ranges of L. sati-

vae and L. trifolii overlap on plants in the families Aster-

aceae, Cucurbitaceae, Leguminosae, Malvaceae, and Solanaceae

(Spencer & Steyskal 1986). Of 40 publications dealing with

Liriomyza as pests of celery (Apium graveolens L.) and

chrysanthemum (Chrysanthemum xmorifolium Ramat.) before

1981, only 4 identify L. trifolii; yet, by 1982, L. trifolii

was recognized as the dominant species in these crops (Lei-

bee 1981b, Trumble 1982). Parrella & Keil (1984) suggest

that this apparent shift likely was due to misidentification

of the two species.

L. sativae is still common on many crop plants. It was

the most abundant species on tomato (Lycopersiscon

esculentum L.) in California during 1981 and 1982 (Zehnder &

Trumble 1984), on watermelon (Citrullus lanatus (Thunb.)) in

Hawaii during 1984 (Johnson 1987), on snap bean (Phaseolus

vulgaris L.) in Louisiana during 1983 and 1984 (Hanna et al.







4
1987), and on tomato (Petitt 1984) and other vegetables in

greenhouses in central Florida from 1982 to the present

(Petitt, unpublished data, 1988).


Biology and Life Cycle of L. sativae

The biology of L. sativae is largely unknown. It is

uncertain whether studies conducted by Oatman & Michelbacher

(1958, 1959) and Oatman (1959a,b, 1960, 1961) on the "melon

leafminer" were conducted with L. sativae because no voucher

specimens exist (Spencer 1981b). Likewise, no voucher spec-

imens exist for the study of L. sativae on tomato by Patel

(1981) and the research was likely conducted, at least in

part, using L. trifolii (D.J. Schuster, personal communica-

tion, 1988). Recently, development and fecundity of L.

sativae has been studied on castor bean, Ricinus communis L.

(Parkman 1987), and voucher specimens have been retained

(J.P. Parkman, personal communication, 1988).

The following description of L. sativae biology was

drawn largely from Oatman & Michelbacher (1958), but each

item described also has been confirmed in the course of the

present study. An L. sativae female lays eggs singly within

the leaf mesophyll by inserting her ovipositor through the

upper leaf surface. Upon hatching, the first instar larva

begins feeding and the larva creates a mine which widens as

the larva develops. Three larval instars occur in the leaf.

The third instar larva cuts a semicircular hole in the upper

leaf surface, emerges from the leaf and drops to the ground

to pupate. Like other cyclorraphous Diptera, L. sativae







5
pupates within the cuticle of the last larval instar, called

the puparium (Comstock 1920). Adults usually mate within

24h of eclosion, and oviposition begins 12-24h after mating

(Oatman & Michelbacher 1958). Females also puncture the

leaf epidermis and feed on plant exudates (Oatman & Michel-

bacher 1958).


Integrated Pest Management for Liriomyza Leafminers

Integrated pest management is needed in crops which

have Liriomvza spp. as a part of the pest complex because,

as previously mentioned, broad-spectrum insecticides have

caused dramatic increases in leafminer populations as a

result of mortality inflicted to their parasitoids. There

is a large complex of ca. 40 species of parasitoids in North

America and Hawaii that attack Liriomvza spp. (Johnson &

Hara 1987), and many other species are active in Europe

(Minkenberg & van Lenteren 1986). Efforts have been made to

conserve naturally occurring parasitoids by the reduction in

number of pesticide treatments (Genung et al. 1978, Waddill

et al. 1981, Petitt 1984), substitution of broad spectrum

insecticides with more selective ones (Phoronezny et al.

1978, Johnson et al. 1980a,b, Parrella et al. 1983), or

elimination of pesticide treatments (Johnson 1987).


Biology of Parasitoids of Liriomvza spp.

The biologies of many species of parasitoids that

attack Liriomvza spp. are unknown; yet, several have been

studied in some detail (see Minkenberg & van Lenteren 1986







6
for a review). Several larval ectoparasitoids in the Eulo-

phidae paralyze the larva, deposit their egg adjacent to it

in the mine, consume the larva and pupate in the leaf among

"pillars" formed from larval meconium. Of these, biological

studies have been conducted with Dilvyphus (=Solenotus)

begini (Ashmead) (Doutt 1957, Allen & Charlton 1981),

Dilvyphus intermedius (Girault) (Hendrickson & Barth 1978a,

Patel & Schuster 1983, Patel 1987) and Dilvyphus isaea

(Walker) (Ibrahim & Madge 1979, Cheah 1987).

Larval-pupal endoparasitoids, however, remain within

the host at the time it leaves to pupate in the soil. Those

which have been investigated include Chrysocharis parks

Crawford (Eulophidae) (Christie & Parrella 1982, 1987),

Ganaspidium utilis Beardsley (Eucoilidae) (Petcharat & John-

son 1988) and Opius pallipes Wesmael and Dacnusa sibirica

Telenga (Braconidae) (Hendrikse et al. 1980). Each of the

parasitoids mentioned above is considered solitary, usually

with only one individual developing per host (Doutt 1964).


Biological Control of Liriomvza spp. in Greenhouses

Unique opportunities exist in greenhouses for the

application of biological control to Liriomyza spp. Few new

pesticides are being developed for greenhouse use, particu-

larly in the United States, because the very high costs

required for development are not likely to be recovered from

the small greenhouse market (Lewis 1977). Manipulations

possible in greenhouses, such as exclusion of certain

insects, e.g., lepidopterous pests, may simplify the pest







7
complex, and may thereby reduce the need for broad-spectrum

insecticides which interfere with parasitoids. The high

value of greenhouse crops enables expenditures to be made

for seasonal inoculative or inundative releases of natural

enemies (van Lenteren 1986). Biological control methods for

other greenhouse pests, such as whiteflies, aphids, thrips

and spider mites, have been developed which are compatible

with each other and with biological control of Liriomyza

species (van Lenteren & Woets 1988).

To date, several greenhouse trials and evaluations have

been conducted with parasitoids of Liriomvza spp. in the

Netherlands (Woets & van der Linden 1982, Westermann &

Minkenberg 1986, Frijters et al. 1986, Minkenberg & van

Lenteren 1987). The leafminer parasitoids Dacnusa sibirica

Telenga and Diqlvphus isaea Walker are being produced and

sold by a company in the Netherlands (Koppert) and a distri-

bution center has been established in the United States

(Koppert & Ravensberg 1986). Trials using parasitoids of

Liriomvza have been conducted in the United States on green-

house-grown tomatoes (Petitt 1984), chrysanthemums (Parrella

et al. 1987), and marigolds (Heinz et al. 1988).


Rearing Opius dissitus

Opius dissitus was first collected in Lake Buena Vista,

Florida, during 1983 (Petitt 1984). It inflicted consider-

able mortality to L. sativae populations in greenhouse toma-

toes, even though no parasitoids had been released (Petitt

1984). Because of this initial success and a very limited








number of options available for controlling L. sativae on

greenhouse-grown food crops, a program was initiated to aug-

ment populations of 0. dissitus in greenhouses at The Land,

EPCOT Center, Lake Buena Vista, Florida (see Robitaille 1983

for description of The Land). Up to this time, taxonomic

descriptions were the only published information on Q. dis-

situs (Muesebeck 1963, Wharton 1984). Research was initi-

ated to adapt and improve rearing procedures that had been

developed for the closely-related parasitoid Opius pallipes

Wesmael in the Netherlands (Hendrikse 1980).

Host instar is known to affect interactions between

parasitoids and their dipterous hosts (Lawrence et al. 1976,

Hendrikse et al. 1980, van Alphen & Drijver 1982); thus,

determining the host instar to be used in rearing was the

first major question to address. Before studies with Q.

dissitus were conducted, the larval instars of L. sativae

needed to be distinguished (Chapter II), and development

times for each instar at different temperatures needed to be

determined (Chapter III). Because host quality is known to

affect parasitoid quality (Zohdy 1976, Price et al. 1980,

Barbosa et al. 1982), studies of intraspecific competition

among host larvae were conducted to determine the maximum

host densities that could be used without affecting host

weight (Chapter IV). Determining the host instar on which

the largest number of "good quality" parasitoids could be

produced was the objective of Chapter V. Earlier studies

indicated that host instar could affect the number







9
(Hendrikse et al. 1980, van Alphen & Drijver 1982, Liu 1984)

and size of parasitoids produced (Liu 1985). Studies with

0. dissitus were then conducted to determine the effect of

host instar and temperature on parasitoid development time

(Chapter VI). These studies will also determine whether

unparasitized L. sativae flies can be separated from

emerging parasitoids by differences in timing of adult

eclosion. Collectively, studies also will determine the

rearing conditions under which life table studies for the

host and parasitoid need to be conducted, so that

demographic principles can be applied to mass-rearing Q.

dissitus (Carey & Vargas 1985, Carey et al., In press).

As studies are completed, results will be used to

improve the ongoing rearing program for 0. dissitus at The

Land. 0. dissitus will be reared and released continuously

to achieve biological control of L. sativae. Experiments to

determine the numbers of parasitoids required to achieve

control are not possible because there is no facility avail-

able in which to conduct this research. Trial and error

will be used to determine whether production levels are ade-

quate to achieve control. Increasing production of Q.

dissitus, however, improves the likelihood of success.

Accomplishment of biological control of L. sativae will be a

significant step toward development of a successful Inte-

grated Pest Management program for The Land.













CHAPTER II
DISTINGUISHING LARVAL INSTARS OF
Liriomyza sativae


Introduction

Liriomvza leafminers cause extensive damage to a vari-

ety of economically important vegetable and ornamental crops

(Parrella & Keil 1984, Minkenberg & van Lenteren 1986, John-

son 1987). Consequently, they are major targets of chemical

(Parrella 1983a, Parrella & Keil 1985, Mason et al. 1987,

Leibee 1988) and biological control programs (Frijters et

al. 1986, Westerman & Minkenberg 1986, Parrella et al. 1987,

Heinz 1988). The likelihood of differential instar suscep-

tibility to insecticides (Parrella et al. 1981, Yu 1983),

and/or parasitism (van Alphen & Drijver 1982) provided the

incentive for this study.

Larval instars of cyclorraphous Diptera generally are

distinguished on the basis of the size and morphology of the

cephalopharyngeal skeleton and the presence and structure of

anterior and posterior spiracles (Bodenstein 1950, Beri

1971, Robinson & Foote 1978, Erzinclioglu 1987). The ante-

rior spiracles are lacking or greatly reduced in first

instar larvae (Bodenstein 1950, Sluss & Foote 1971, Lawrence

1979). Morphology of posterior spiracles may be used to

separate larval instars in some species (Robinson & Foote

1978). Although larval instars of some species can be






11

differentiated using spiracles, the cephalopharyngeal skele-

ton or its parts are most often used.

The cephalopharyngeal skeleton consists of the mouth

hooks, the hypopharyngeal sclerite and the tentoropharyngeal

sclerite (Teskey 1981). The mouth hooks consist of the

right and left mandibles which are united basally and work

as a unit (Frick 1952, Teskey 1981). The morphology of

mouth hooks can be used to distinguish larval instars in

some species (Bodenstein 1950, Robinson & Foote 1978, Law-

rence 1979). The most evident morphological distinctions

are based on dentition. For example, numbers of teeth

increase from 1 in first instar drosophilids, to 2-3 in sec-

ond, and 9-12 in third instar larvae (Bodenstein 1950). In

addition to, or as an alternative to, the use of mouth hook

morphology, the length of the cephalopharyngeal skeleton has

often been used in differentiation and has been a consistent

and reliable character (Beri 1974, Ipe 1974, Hendrickson &

Barth 1978b, Lawrence 1979, Hendrickson & Keller 1983).

On the basis of the cephalopharyngeal skeleton, three

identifiable instars have been reported among the cyclorrap-

hous Diptera (Bodenstein 1950, Lawrence 1979, Erzinclioglu

1987). One characteristic of members of this group is the

formation of a puparium from the integument of the third

instar (Comstock 1920). The cephalopharyngeal skeleton is

shed after the last larval instar retracts its head into the

puparium. Thus, the third instar is the last actively feed-

ing instar before the pupal stage. Although Parrella (1987)








stated that a fourth instar larva occurs in Liriomyza, the

present study is concerned only with distinguishing the

three actively feeding phytophagous larval instars.

Larval instars in this study were distinguished using

measurements of either the length of the mouth hooks or the

combined lengths of the hypopharyngeal and tentorpharyngeal

sclerites (hereafter referred to as the cephalopharyngeal

skeleton). Comparisons of sizes and growth ratios of these

structures were made with other Liriomvza species. Data

were examined to determine whether the Brooks-Dyar Rule of

geometric growth could be applied to larval Agromyzidae

(Hutchinson & Tongring 1984, Daly 1985).


Materials and Methods

Experimental Procedure. Bush lima beans (Phaseolus

limensis var. limenanus Bailey 'Henderson') were grown in a

greenhouse in 15cm diameter pots at two plants per pot in a

steam-pasteurized peat-vermiculite mixture (Speedling Inc.,

Sun City, FL). Plants were watered with a modified Hoagland

solution. (See Appendix A for composition of nutrient solu-

tion.) Lima beans 10-14d old were placed in a 1.9 x 1.1 x

0.6m cage containing a colony of adult L. sativae and ovipo-

sition occurred for 6h. Plants were held in environmental

chambers at 20 or 2510C and a 14h photophase beginning at

0700 hours.

Larvae were collected from primary leaves every 6 and

12h at 25 and 200C, respectively, beginning the third day

after oviposition. Samples were subsequently preserved in a








70% ethanol:water mixture. Larvae were cleared in a 10%

solution of potassium hydroxide prepared with 70% ethanol

for 2-12h, depending on larval size. The lengths of mouth

hooks (Fig. 2.2) and the cephalopharyngeal skeletons (Fig.

2.2) of 560 larvae were measured on a compound microscope at

400x using an ocular micrometer. Measurements were made to

the nearest micrometer unit which was equal to 2.51 pm.

Light micrographs were taken using a 4x5 in. camera attached

to the microscope.

Data Analysis. The length of mouth hooks for each lar-

va was plotted with its cephalopharyngeal skeletal length to

show the degree of separation between instars. The natural

logarithms of the measurements were regressed against pre-

sumed instar number (independent variable) to confirm that

no instars were missed (Daly 1985). Growth ratios were cal-

culated by dividing the size of a structure in an instar by

the size of the equivalent structure in the preceding

instar. Constant growth ratios were expected according to

the Brooks-Dyar Rule (Hutchinson and Tongring 1984).

Voucher Specimens. Specimens of adult L. sativae were

sent to K. A. Spencer (Univ. of Exeter, Exwell Farm, Bray

Shop, Callington, PL17, 8QJ, Cornwall, United Kingdom) and

their correct identification was verified. Voucher speci-

mens were deposited in the Florida State Collection of Arth-

ropods in Gainesville, Florida.








Results and Discussion

A bivariate plot of lengths of mouth hooks and cephalo-

pharyngeal skeletons showed clear separation among the three

larval instars (Fig. 2.1). The absence of abrupt deviation

in linear regression of the natural logarithms of the length

of mouth hooks and cephalopharyngeal skeletons on the pre-

sumed instar number (r=0.99 for both) indicated that no

instar was missed. The ranges in size of these structures

in the three instars do not overlap (Table 2.1). The mean

lengths of the cephalopharyngeal skeletons were almost the

same as the mean lengths of 0.09, 0.15, and 0.23 mm reported

for first, second, and third instars, respectively, by Oat-

man and Michelbacher (1958) for the "mouth hooks" of L.

pictella (probably L. sativae, see Spencer 1981b). These

authors apparently used the term "mouth hooks" to describe

the cephalopharyngeal skeleton.

The mouth hooks of second and third instar larvae are

morphologically similar, but the first instar is distinctly

different (Fig. 2.2). The second and third instar mouth

hooks were less than one-half as wide as long (including

teeth) in side view and had two distinct pairs of alternat-

ing, ventrally curved, terminally-pointed teeth. The right

side of L. sativae third instar mouth hooks was longer than

the left in front view, as is usual among larval Agromyzidae

(Frick 1952, Beri 1983). The relative size of each mandible

was not compared in earlier instars. The first instar mouth

hooks were one-half or more as wide as long and only a pair


























I I~



r*jt
*


*
*X

*
II ] *

*c **
**


*


Int*




^*
*X


CD Lf 1


SIOOH qL4.noW


Cr,

J0


0 0

L46UB-1


cC
-o
-o












c-



-C
c.





CJ
ai
-o







a


00
- I







C2


-C
-a)

o




'4-



_J


r-
(oIJ


0
1I
O





0
0
(o



4)
0
0a
in
o
o






04


e.


-0


U) 00
>-jo


4) c)

0 0

'4-










t;4

.rqi
F43
Ul (












*r
















0 I I I



4 0 03
C 0 1 1 1



r 000
QO 4) 00 o
H H N ON
r-4 -4 0 'H H-
0 4 *


4 0 0 0

o. 0 0 0 >1
0
V 4) 1
0 o o o (a
H 0 I0 a m
0 (0 H m Cc
0x H N 0

,C o o o'



1 W4J
Cu .



0 0 0 0 >a
M 0

O *

f.4 m 1 L CO 4
00 0 0




O 0 0 H
P O Q)
Hc I I U
10i r in in coi T







H c o o o go
S0 0 0 M



U U) 0 0 0 a)
0 0 I

0C Vi
S4). p




44 0 0 0 WQ
0 0 0



0 ( 0 0 050
0l 1 3 0 0 0 HO


0 0 0 r.


ON H H U)
)(4-4 >1>
0P p








E--
































Fig. 2.2. Cephalopharyngeal skeleton including mouth
hooks of first (A, B), second (C, D), and third
(E) instar larvae of L. sativae. The presence
of partially formed mouth hooks (arrows)
posterior to the functioning mouth hooks in B
and D indicate the onset of ecdysis to the next
instar. Length of mouth hooks (L-M), and
length of the cephalopharyngeal skeleton (N-0),
as measured in this study, are shown in E. The
same magnification was used in A-E.






18



s,4Pkr


A. z ,





B.


~r1









C.









D.









E.


0. Imm








of apical teeth were apparent. Occurrence of partially

formed mouth hooks along the hypopharyngeal sclerite indica-

ted the onset of ecdysis to the next instar (Fig. 2.2B and

2.2D).

As in other agromyzids (Beri 1973), first instar L.

sativae apparently were metapneustic, having only the poste-

rior pair of spiracles. The anterior spiracles of second

and third instars are stalked with numerous spiracular

bulbs. The posterior spiracles have three bulbs, one of

which is longer than the other two, as is typical of

Liriomvza (Beri 1973).

Growth ratios of cephalopharyngeal skeletons between

first and second instar L. sativae were larger than between

second and third instar larvae (Table 2.2). In L. sativae,

L. pictella, L. trifolii, and L. brassicae, all of which

feed on leaf mesophyll tissue, the ratios were consistent,

with the exception that ratios of cephalopharyngeal skeleton

length between first and second instar L. brassicae were

noticeably larger than the rest. Fewer data were available

for mouth hooks. However, the growth ratio between second

and third instar mouth hooks was very consistent and close

to the ratio observed for the cephalopharyngeal skeleton

(Table 2.3). The constancy of the growth ratio may be

related to the similar mode of feeding of these species

(Hutchinson & Tongring 1984).

Although measurements of the mouth hooks and cephalo-

pharyngeal skeletons in this study were made on cleared,


















0

1-

U( >
4)
0rj



>1
0







CO
40-
00


04X


a-9-
C0 4
0 0





0)0M
00 4

tiC



oO
X 0






0 4

4) 4



04.)


.! 0
'PU

0 10




O ca
*O
* *)

4) .4
0
4: 0)

0 co
0 (a





X)

to


co
Ln

CO


S009


0) 4-)

.-4<
U O



r-4 0%


0 0co
(0 ja


4-4

0 0
0 (0
04 C

4 n
P(44-H



























*.M

-H


to

0
0 N
D>1
r:


O0H


CO
0 *)
00






3
O4 0













C,(
0 'V









MO
( 4













*-


11
I
0






























I -











4H 4)
4 *P




-1-1


I O
f-I




























co







CO\j
u-I i-











N Ny

(u c~








slide-mounted larvae under a compound microscope, larval in-

stars also can be distinguished by measurements of these

structures under a dissecting microscope. At 75x total mag-

nification all instars could be differentiated by measuring

the length of the cephalopharyngeal skeleton. Because the

body of the later instars becomes more opaque with time in

ethanol, thereby obscuring the cephalopharyngeal skeleton,

measurements were easier with freshly collected larvae. If

the cephalopharyngeal skeleton is not entirely visible, sec-

ond and third instar larvae can be separated using measure-

ments of mouth hooks alone (150x total magnification suffi-

cient).


Conclusions

Larval instars of L. sativae can now be distinguished

on the basis of lengths of the mouth hooks and/or cephalo-

pharyngeal skeletons. Rapid instar determination on freshly

collected specimens under the dissecting microscope makes

immediate processing of samples possible. To determine when

all individuals have reached a given instar, cohorts can be

subsampled frequently, if necessary, using this technique.

Computer-assisted measurement may further facilitate use of

this technique (Daly 1982).

Instar determination by measurement of the cephalo-

pharyngeal skeleton or its parts would likely be useful for

most members of the Cyclorrhapha (=Muscomorpha), because the

cephalopharyngeal skeleton is characteristic in this group






23

(Teskey 1981). The structure is not present in the Nema-

tocera or orthorrhaphous Brachycera (Teskey 1981).













CHAPTER III
EFFECTS OF TEMPERATURE ON DEVELOPMENT
OF IMMATURE STAGES OF Liriomyza sativae


Introduction

Detailed biological and ecological studies on L.

sativae are lacking (see Parrella & Keil 1984, Parrella

1987). Of the many studies on temperature-dependent devel-

opment in Liriomyza spp., only Parkman (1987) definitely

worked with L. sativae. Oatman and Michelbacher (1959)

probably also studied L. sativae (see Spencer 1981b), but

absence of voucher specimens makes positive identification

impossible. Of all previous studies on development in

Liriomyza, only three have measured the effects of tempera-

ture on development of larval instars (Webb & Smith 1969,

Hendrickson & Keller 1983, Beri & Chandra 1983).

The inability to distinguish larval instars of

Liriomyza spp. and to predict their development rates likely

will delay progress toward successful biological control

programs, particularly where parasitoid mass-rearing is a

component of the program. Considerable evidence exists with

dipterans that host instar may influence many aspects of

host-parasitoid interactions, including host detection and

host acceptance (van Alphen & Drijver 1982), parasitoid sur-

vival (Lawrence et al. 1976, van Alphen & Drijver 1982) and

developmental synchrony between host and parasitoid








(Pemberton & Willard 1918, Lawrence 1982). Within the genus

Liriomyza, host age has been demonstrated to affect the time

required to find hosts (Hendrikse et al. 1980) and number of

hosts parasitized (Lema & Poe 1978). Some parasitoids of

Liriomyza spp. are also known to have specific host instar

preferences (Hendrickson & Barth 1978a, Johnson et al.

1980c).

A predictive model for development of L. sativae larval

instars would be very useful in studies of host-parasitoid

interactions. Because it would be useful to predict when

all individuals in a cohort have reached a given instar, the

variability in development times of individuals within a

cohort must be simulated. If time required for development

is considered a delay before the next stage can be reached,

then the duration of the delay varies for different individ-

uals. That is, passage of individuals into the next stage

is distributed over time. Distributed delay models are

appropriate for this purpose (Manetsch 1976). To avoid the

complexity of a delay that changes duration with tempera-

ture, the duration of the delay can be stated in degree-day

units (Welch et al. 1978, Gutierrez & Baumgaertner 1984).

In this study, a distributed delay was used to model devel-

opment of L. sativae using data from constant temperature

experiments. The predicted timing of developmental events,

such as egg eclosion and larval molts, was then compared to

development observed under fluctuating temperature condi-

tions.








Materials and Methods

Experimental Procedure. Bush lima beans were planted

and maintained as described in Chapter II. Lima beans 10-

14d old were placed in 1.9 x 1.1 x 0.6 m cage containing a

colony of adult L. sativae and oviposition occurred for 6h

beginning 0900 hours. Plants were held in environmental

chambers for the remainder of the experiment at constant

temperatures of 20, 25, 30 or 3510C and a 14h photophase

beginning 0700 hours. A low temperature (150C) treatment

was included when to address the feasibility of cool storage

of pupae, which may be useful during times of abundance in

mass-rearing. A CR7 datalogger (Campbell Scientific, Logan,

UT) monitored temperatures each minute and maintained

running averages.

Eggs were observed under a stereomicroscope using

transmitted light every 6h to determine when eclosion

occurred (50x total magnification). At temperatures above

200C, eight or more larvae were removed from leaves every 6h

and preserved in a 70% ethanol:water mixture. Larvae devel-

oping at 200C were collected every 12h. This longer inter-

val between observations was sufficient because development

was greatly retarded at 200C. Larval instars were differen-

tiated on the basis of measurements of mouth hooks and the

cephalopharyngeal skeleton (see Chapter II). The percentage

of larvae in each instar was determined for each 6h

collection.








For purposes of this experiment, the end of the third

instar was considered to be the time when larvae exited from

the leaves. Larval emergence from the leaf signaled the end

of the larval stage since pupariation ensued shortly there-

after. Preliminary experiments with the same photophase

showed that 96% of larvae emerge daily between 0700 and

1400, with the largest percentage emerging between 0800 and

1000. In these experiments, emerging larvae were collected

only once per day (after 1400) and the emergence time of

0900 was used in all calculations.

Development time of pupal stage L. sativae was studied

in experiments separate from those addressing larval devel-

opment. Plants were maintained as above; however, develop-

ment time was determined for L. sativae pupae which survived

exposure to parasitism by Q. dissitus. Time from L. sativae

oviposition to larval emergence and to adult eclosion was

recorded.

Analysis of Development Rates. The average time

required a) from the midpoint of the oviposition period to

egg eclosion, b) from oviposition to completion of each molt

and c) from oviposition to larval emergence from leaves was

calculated. Since larvae were removed from the leaf and

killed to enable measurement of the cephalopharyngeal skele-

ton, no single individuals were followed through the entire

experiment. Consequently, standard errors of development

times could not be calculated for larval stages. The mean

time of completion of each molt and the variance were








calculated from percentages of larvae in each 6h sample that

had completed the molt. Stadial durations were calculated

by subtraction of the times required from oviposition to

complete successive molts. For example, the duration of the

second instar was determined by subtracting the average time

of completion of the first molt from the average time of

completion of the second molt.

Linear regression was used to relate development rates

to temperature. The hypothesis that a regression coeffi-

cient was different from zero was tested using F-test within

the SPSSX Scattergram procedure (SPSSX 1986). The hypothe-

sis that regression coefficients differed from one another

was tested using the F-test described by Sokal & Rohlf

(1981). The lower developmental threshold (LDT) was calcu-

lated as the x intercept of the regression line. The number

of degree-days (DD) required a) from oviposition to egg

eclosion, b) from oviposition to completion of each molt,

and c) from oviposition to larval emergence from the leaf

were calculated as follows:

DD=(T-LDT)DEL,

where T was the average temperature in the environmental

chamber and where DEL was the average time in days required

for development.

Modeling and Validation. Development times of eggs and

larvae were modeled using a time-invariant distributed delay

(Manetsch 1976). Eggs deposited during the 6h oviposition

interval are considered an impulse input occurring at the








midpoint of the interval and the transit times of individu-

als through the delay (development times) are described by

an Erlang density function given by:

y(t) = exp(-kt/DEL) (k/DEL)k t(k-1)/((k-1)!)

The observed mean (DEL) and variance (VAR) in development

times (expressed in degree-days) determine the parameter 'k'

(k = DEL2/VAR), which specifies the density function. The

cumulative number of individuals emerged by a given degree-

day is calculated by multiplying the area under the

probability density function by the number of individuals

being processed through the delay. A computer program was

written in BASIC which calls for the mean and variance in

development times (expressed in degree-days) as input,

performs these calculations and produces an output file in

which the number of individuals emerging is given for each

degree-day accumulated (see Appendix B).

An experiment was conducted to validate the model. The

method was the same as described for egg and larval develop-

ment experiments, except that temperature fluctuated between

3310C during the 14h photophase and 2710C during the

scotophase. Eclosion of eggs was observed and 20 larvae

were collected at intervals of approximately 4DD. The Kol-

mogorov two-sided goodness-of-fit test was used to compare

model predictions for egg eclosion, molting, and larval

emergence with observed development (Conover 1980).

Voucher Specimens. Voucher specimens of L. sativae

were collected and submitted as described in Chapter II.








Results

Development. L. sativae egg development time decreased

from ca. 4.7 to 1.7d as temperature increased from 19 to

340C, respectively (Fig. 3.1). (See Appendix C for data on

development of immature stages of L. sativae.) Egg develop-

ment rate increased linearly with temperature up through

3000C, but more slowly thereafter (Table 3.1).

Larval development time declined from ca. 7d at 200C to

3.4 d at 340C, respectively (Fig. 3.1). The increase in

development rate of larvae with increasing temperature was

linear up through 340C (Table 3.1).

Pupal development rate increased linearly up through

300C, but did not increase with further increases in temper-

ature (Fig. 3.2, Table 3.1).

The development rate of first and second instar larvae

increased linearly with temperature (Fig. 3.3). Slopes of

the lines for first and second instars differed signifi-

cantly from each other (F=11.80; df=1,20; P<0.005). This

indicates, for instance, that the development rates of

second instar larvae change more rapidly with temperature

than do rates of first instar larvae. Linear regression of

development rates of third instar larvae on temperature was

significant (P<0.05), but only 30% of the variability could

be explained by temperature.

Development rate from oviposition to completion of

either molt or larval emergence increased linearly with
























9 ]-------------------7 -- 60
9 O B
o -Egg 8 D
co, o 0
00
0 C
v on Larva o

E O

0



(4 *


Of
E I m0


0 E '






10 15 20 25 30 35

Temperature (oC)
Fig. 3.1. Relationship of development time and devel-






bean.
0 0
10 15 20 25 30 35

Temperature (C)




Fig. 3.1. Relationship of development time and devel-
opment rate to temperature for the egg and
larval stages of I. sativae on bush lima
bean.





















Table 3.1. Linear regression equations for relationships
between development rate and temperature for
the egg, larval and pupal stages of L.
sativae on lima bean primary leaves.

Temps. (OC)
Life omitted from
Stage n regressiona Equationb,c r LDT(OC)d

Egg 9 >30 y=3.04X-36.59 0.99 12.0

Larva 12 None y=1.02x-5.24 0.98 5.1

Pupa 7 >30 y=0.72x-7.40 0.99 10.3

a Temperatures beyond the linear portion of the curves
were omitted.
b x=Temperature in C.
c y=Development rate (1/Days x 100).
LDT=Lower developmental threshold (x intercept).

























35 20 /-


:D a X
m0 30 0


E C



01 -4---
E 15 -
0
10 -0
> 0 0




0 I
10 15 20 25 30 35

Temperature (CoC)




Fig. 3.2. Relationship of development time and devel-
opment rate to temperature for the pupal
stage of L. sativae reared on bush lima bean.
























o /0 r=0.95
S120. + 1st Instar
\ a 2nd Instor
C
o 100- 3rd Instor o r=O.97


080
0 *


S/.. ,J---r- =0.65
tl 800


- 40 -
a <1+
0
C

10 15 20 25 30 35 40

Temperature COC)




Fig. 3.3. Relationship of development rates to
temperature for larval instars of L.
sativae reared on bush lima bean.








temperature (Table 3.2). The slope of each line was

significantly different from zero (P<0.0001). The lower

developmental thresholds (LDT's) for these regressions

ranged from 7.5 to 12.20C, but a standard LDT of 100C was

used for calculation of the mean and variance in number of

degree-days required for development, as suggested by Pruess

(1983).

Modeling and Validation. The distributed delay program

used the observed mean and variance required for development

in constant temperature experiments to determine the cumula-

tive number of individuals that should reach a stage or

instar by a given time (in DD) (Fig. 3.4, solid lines, x

axis scaled so the 15DD intervals represent one calendar day

at 250C). Egg eclosion and timing of molts observed under

fluctuating temperatures agree with the model; whereas,

larval emergence from leaves does not agree (Ho: Observed=

Predicted; T=0.40, n=4, P>0.20 for egg eclosion; T= 0.51,

n=6, P>0.05 for molt to second instar; T=0.56, n=5, P>0.10

for molt to third instar; T=0.64, n=5, P<0.05 for larval

emergence). The observed number of degree-days required for

50% of the individuals to have advanced to the next instar

agrees very well with the model. However, the variance in

model output is greater than observed in this experiment,

particularly for individuals molting to the third instar.

The disparity between observed larval emergence and

model output was due primarily to the fact that larval

emergence was actually discontinuous in the 14L:10D regime,





















ao
I 10
a(.0
) 0

0)
tO
C
c c



g 4
a)
0P 0


C A



0
4) 0
4 0*tP
> a-









>k 0




D-4 4.)
4 0













a)
4H H
0 e
moO






0
3 1



S0











aC


0 c
U
0
0
A


Q IX



















4
SIo











C
O
-P


&
M


M N
N m
* *





co a
H co
* *
co H
co H-i
1-4






N L





a\ a
01 0





0 0




* *
o o




X N
H-
I II
x x

H*N
H 0)








02


> k
rC S


U
o
.--

.4
a)

0



.4 0



II 0 0M
S0a)
Xo 5





(a 41
OH a)
00
M 0
5.4

>irP Q)
0 4.)
vH 0
S(r-4

140 H
Ho) o
e CO

a)V 0








a) a)






0 O0
















I I I I I I I I I I I


+


-~ -- I


n +
+

+ +



+ ++





4+
(+n



+ ~ L
t"


0 *P


a -4 r-
> p1

0) ) 0





o a0
a X





A w
D M 4-1 -4
a 4 r -1 14
L 0 0 0 3
0) ) 0
0094








00
0 C iO 0
0 ( 4 *H 0













Ua
H *V4 0







CD C C-
1 1M "1

0 uo


Q OO W
e > r.i




0 w0

in CO 0 C ON

0 + In e
o ^0


J4SUT / 9be4s 04 p8OueAPV % 'un3








while the model allowed for continuous emergence. In this

photoperiod no larval emergence was observed after 1600

hours. If model output were modified so that the larvae

which were to emerge after 1600 hours, instead, all waited

and emerged during the first hour of the next photoperiod,

it would then agree very well with observed data (Fig. 3.4,

T=0.17; n=5; P>0.20)


Discussion

Development. It is possible that L. sativae egg devel-

opment rate did not increase with increases in temperature

above 300C because of differences between ambient air tem-

perature and leaf temperature, as suggested by Leibee (1984)

for L. trifolii (Burgess). If the non-linearity resulted

because of these temperature differences, however, larval

development also should show deviations from linearity at

temperatures above 300C. However, no substantial deviations

from linearity in larval development rate were observed at

350C in either study. It seems likely that ambient relative

humidity would be an important factor in determining the

extent to which a leaf could maintain its temperature below

ambient by evaporative cooling. The extent to which this

phenomenon occurred at the 8010% relative humidity in these

experiments is unknown.

The development time of L. sativae larvae was very sim-

ilar to that determined for L. trifolii on tomato (Schuster

& Patel 1985) and for another Liriomvza species on bush lima

bean (Oatman & Michelbacher 1959) (probably L. sativae, see








Spencer 1981b) (Fig. 3.5). The much longer larval develop-

ment times for L. trifolii on celery (Leibee 1984) presum-

ably are caused by differences between host plants.

Development times of larval instars have been studied

in Liriomyza trifolii (Burgess) by Webb & Smith (1969)

(reported as L. munda), in L. brassicae (Riley) by Beri &

Chandra (1983), and in L. trifoliearum by Hendrickson & Kel-

ler (1983). The latter study involved only two temperatures

and will not be discussed in detail. As comparisons are

made to these other two studies, it should be emphasized

that differences in methodology exist and these differences

are confounded with differences that may exist between the

species. Despite these difficulties, development of larvae

of these Liriomyza species was compared.

The mean percentages of the total egg-larval develop-

ment time that was spent in each stage or instar were quite

similar between L. sativae and L. brassicae (Table 3.3).

In L. trifolii, however, the egg stage was proportionately

longer and third instar shorter than in the other species.

When experiments with L. sativae and L. brassicae were run

at different temperatures, the percentage of time spent in

each instar deviated only slightly from the mean in all

cases except the third instar. In these species, however,

the percentage of the total egg-larval development time

spent in the third instar increased with increasing tempera-

tures (Table 3.3). This may be explained in part by the























25---
Ls Lima Bean

20-
o Lt Celery


Temperature (oC)


Fig. 3.5.


Larval development time for Liriomyza
species on various host plants. (Ls = L.
sativae, Lt = L. trifolii.)


o Ls? Lima Bean


o Lt Tomato


a


b *0







41




0

0 0 0
V0 o o o
IA r o in
H .4 m nO N
0C 0 0
to 0 $4 4 o o4
)Q) 4o4
C= B U U U
0t 0 0 0
Vj tr 0 In In
0 *- -
ar 4
X) Ix m In %D

M H r- rr




4 O

40 c0
40 0 %0 o m
4-1 0)0CO






00 0 N
1I 4-)
04i1 0 iI 4 .






4) 0 (D m 01

10$4 o





M. .





O I -
4Jam m ni 4 0 0 e4





o o 0








Qtr M
14J 41 Cn C;










* *M
0 M M 0- -
+> 4 H al in *n






* 4 0 .
.H 0 0
4 *4 *








o .I c4

0 0
10441 tP %a .'
S14 IX.




0 V




4) 0t -"I
M 10 $4 i14
> M








fact that emergence of the third instar from the leaf occurs

only at certain times of day. Because the duration of the

third instar is near Id at 300C, the "gates" or "allowed

zones" when emergence can occur (sensu Saunders 1982) may

limit the extent to which this stage can be shortened at

higher temperatures.

Development times for L. sativae pupae which had been

reared on bush lima bean in this study were very similar to

those reported for L. sativae on castor bean (Parkman 1987)

and L. trifolii on celery (Leibee 1984).


Conclusions

A predictive model for development of larval instars of

L. sativae was developed and validated using two different

methods. The table reporting the predicted percentage of

immature L. sativae advanced to the next stage or instar for

each degree-day accumulated can easily be used with a hand-

held calculator to answer "what if" questions. For example,

the temperature for rearing which will result in a given

stage or instar at a particular time can be determined.

This was used for conducting experiments with particular

larval instars in Chapters V and VI. It also can be used to

determine how rearing must be conducted to satisfy a set of

constraints (see Appendix B, Applications of Larval

Development Model).

The model was developed using data from cohorts which

resulted from a 6h oviposition beginning 0900 hours. These

starting conditions were used to reduce variability in








molting times among numbers of a cohort and to simplify the

modeling process. The short oviposition interval approxi-

mated the impulse input assumed by the distributed delay.

Distributions of development times can be predicted only for

these starting conditions. To generalize the model so that

it will work under a variety of starting conditions would be

difficult. How changes in the length of the oviposition

interval would affect the variance in development times is

not known and would need to be determined by experimenta-

tion.

Modification of the model to allow larval emergence

only during certain hours in the photoperiod would be diffi-

cult in unpredictable environmental conditions because the

accumulations of DD (on which the model is based) would not

correspond in a predictable way with the photoperiod. For

rearing, however, it will be necessary to determine the day

on which first larval emergence will occur.

The distributed delay model should have general utility

for distributing development of individuals of any species

for which development is continuous and the mean and vari-

ance in development are known. It was used here with

degree-day time units so that development under fluctuating

temperatures could be modeled. For use in this manner, the

assumptions of the degree-day approach must be met. The

model would also be useful to distribute development of a

cohort at a constant temperature over calendar time.













CHAPTER IV
INTRASPECIFIC COMPETITION AMONG Liriomyza sativae
LARVAE IN PRIMARY LEAVES OF LIMA BEAN


Introduction

Competition occurs when organisms use a resource that

is in short supply or limit access to a resource, thereby

reducing its availability to others. Individuals may com-

pete for the use of a limiting resource without interacting

aggressively (exploitative competition), or an individual

may directly or indirectly limit another's access to a nec-

essary requirement or resource (interference competition)

(Birch 1957). As a consequence of these competitive inter-

actions, survival or fecundity of members of one or more

genotypes or species may be affected.

Competition among individuals of the same species

(intraspecific competition) has resulted in reductions in

survivorship and/or larval weight in lepidopteran (Murai

1974, Bultman & Faeth 1986), hymenopteran (Tuomi et al.

1981), and dipteran (Oatman 1960, Parrella 1983, Quiring &

McNeil 1984a, Stiling et al. 1984, Potter 1985) leafmining

species. With leafminer flies in the family Agromyzidae,

larval weight and/or size reductions have subsequently

resulted in decreased fecundity and longevity of adults

(Parrella 1983, Quiring & McNeil 1984b). Since the goal of

this research is to improve rearing techniques for both








Liriomyza sativae Blanchard and its endoparasitoid, Opius

dissitus Muesebeck, weight or size reductions in host larvae

are also of concern for their potential impact on para-

sitoids being produced. Smaller host insects often produce

smaller parasitoid females (Arthur & Wylie 1959, Sandlan

1979, Liu 1985, Opp & Luck 1986), which may be less fit

because of reductions in longevity and/or fecundity (Sandlan

1979, Liu 1985, Opp & Luck 1986). Because intraspecific

competition among larvae may reduce survival and/or repro-

ductive performance of adult leafminers, the demographics of

rearing will vary with the extent of competition. Determin-

ing the extent to which such competition occurs was consid-

ered a priority in the effort to improve rearing techniques

and a prerequisite to further detailed research on host-

parasitoid interactions.

Intraspecific competition was studied among L. sativae

larvae in primary leaves of bush lima bean. The objectives

were to evaluate the effect of larval density on competitive

interactions among L. sativae in primary lima bean leaves

and to determine the maximum larval density at which L.

sativae could be reared without negatively influencing pupal

weight or pupal survival.


Materials and Methods

Experimental Procedure. Bush lima beans were planted

and maintained as described in Chapter II. Lima beans 10-

14d old were placed in a cage containing a colony of adult

L. sativae to allow oviposition for ca. 4h on primary








leaves. Plants were maintained thereafter in an environmen-

tal chamber at 2510C and 8510% relative humidity with a

14h photophase beginning 0700 hours.

Within 4d after oviposition, L. sativae eggs hatched

and the resulting larvae had created small mines. Using a

dissecting microscope, the numbers of first instar larvae

per leaf were determined. Interactions among the larvae

were observed and recorded daily. Cannibalism was consid-

ered to be the cause of death when the act itself was

observed or when a larva was found dead at the intersection

of two or more leaf mines. Larval deaths due to other

unknown causes were also observed and recorded.

On the seventh day after oviposition (day 7) the leaves

were excised and their petioles placed in a modified Hoag-

land nutrient solution contained in a water reservoir (Syn-

dicate Sales, Inc., Kokomo, IN). Leaves were suspended over

0.5 liter clear plastic cups into which emerging third

instar larvae dropped. Plastic cups were replaced and

puparia counted at ca. 1500 hours daily. Preliminary exper-

iments showed that by 1500 hours all larval emergence had

occurred for the day. Puparia were held in an environmental

chamber until they were weighed 5d later on an analytical

balance (Mettler Instrument Corp., Hightstown, NJ, Model

H54AR). Not all puparia were weighed because many were

stuck together in groups that could not be separated without

damaging them. A Li-CorR 3000 leaf area meter (Li-Cor,








Lincoln, NE) was used to measure leaf areas after all larvae

had emerged. Thirty-one leaves were used in the experiment.

The area of individual leaf mines was determined by

placing a leaf under a video camera and tracing the perime-

ter of the mine on a calibrated digitizing pad (Southern

Micro Instruments, Atlanta, GA). The portion of the mine

made by the first instar could not be included because it

was too narrow to trace accurately. Failing to include area

mined by first instars should not greatly affect estimates

of total leaf area mined, since observations of the closely

related species Liriomyza trifolii (Burgess) showed that

area mined by a first instar larva was only 4% of the total

mine area (Parrella & Bethke 1988).

Data Analysis. The number of larval deaths by canni-

balism was regressed on first instar larval density. The

percentages of larvae killed by cannibalism and other causes

were calculated by dividing the number of deaths by the num-

ber of first instars determined at the initial count. These

percentages were regressed on first instar larval density.

Wilcoxon's signed-ranks test was used to compare magnitude

of mortality from cannibalism and other causes (SPSSX 1986).

Egg-larval development time, pupal weight, and the per-

centage of puparia from which adults closed were regressed

on third instar larval density. Third instar density was

used as the independent variable in these instances because

density had been reduced by cannibalism during earlier

instars, and exploitative competition that may affect weight








or development would likely occur during the third instar,

when most leaf consumption occurs. The SPSSX Scattergram

procedure was used to calculate correlation coefficients (r)

and to test the significance of regression lines (Ho:

slope=0).

Pupal weights of larvae that emerged on different days

and pupal weights of low density and high density groups

were compared using the F-test within the SPSSX Oneway Pro-

cedure (SPSSX 1986). Mean separation and homogeneity of

variances were tested using the LSD test and Bartlett's Box

F-test within the Oneway Procedure (SPSSX 1986), respec-

tively. When unequal variances were encountered, Brown and

Forsythe's modified F-statistic was used to approximate the

critical region of the standard F-statistic when variances

are equal (Gill 1978).

Voucher Specimens. Voucher specimens of L. sativae

were collected and submitted as described in Chapter II.


Results

Larval survivorship between the first instar and the

time of larval emergence from leaves was an average of 89.6%

unweightedd mean; n=1689). Percentage larval survivorship

declined with increasing larval density (r=-0.47; P<0.01)

(Table 4.1). Cannibalism was the only identifiable source

of mortality, and from 0 to 11.3% unweightedd mean=5.7%) of

larvae died as a result of cannibalism which was directly

observed. Loss of larvae in crowded areas of leaves between

daily observations was also likely due to cannibalism










Table 4.1. Influence
mortality


of L. sativae larval density on
in primary leaves of lima bean.


No. 1st % Mortality
Instar Density % Larval
Larvae (larvae/cm2) Cannibalisma Lostb Otherc Survivald


25
18
32
27
38
47
32
30
39
32
52
46
44
50
42
52
57
58
50
60
61
59
55
75
75
53
72
80
92
105
131


0.23
0.23
0.31
0.32
0.41
0.43
0.43
0.46
0.47
0.50
0.55
0.56
0.57
0.60
0.63
0.64
0.67
0.67
0.68
0.69
0.73
0.76
0.77
0.77
0.83
0.83
0.89
0.90
1.14
1.36
1.74


8.0
5.6
0.0
7.4
2.6
8.5
3.1
3.3
7.7
0.0
3.8
8.7
6.8
2.0
7.1
1.9
5.3
5.2
8.0
6.7
6.6
5.1
3.6
6.7
8.0
3.8
2.8
11.3
6.5
5.7
6.1


0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
9.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
3.4
0.0
0.0
0.0
3.4
0.0
0.0
0.0
7.5
0.0
0.0
3.3
4.8
5.3


4.0
5.6
12.5
0.0
0.0
2.1
0.0
0.0
0.0
0.0
0.0
2.2
9.1
0.0
7.1
1.9
3.5
3.4
0.0
5.0
1.6
0.0
1.8
2.7
2.7
5.7
4.2
1.2
4.3
1.9
9.2


88.0
88.8
87.5
92.6
97.4
89.4
96.9
96.7
92.3
90.6
96.2
89.1
84.1
98.0
85.8
96.2
91.2
88.0
92.0
88.3
91.8
91.5
94.6
90.6
89.3
83.0
93.0
87.5
85.9
87.6
79.4


a Larvae which died as a result of observed cannibalism.
Numbers of deaths due to cannibalism increased with
larval density (r=0.73; P<0.0001), but percentage of
larvae dying from cannibalism did not increase with
larval density (r=0.15; P>0.40).
b Larvae were lost between daily observations in crowded
areas of leaves.
c Larval death was observed, but cause not known.
d Larval survivorship decreased with increasing density
(r=-0.47; P<0.01).








unweightedd mean=1.5%), but these larvae are not included in

further references to cannibalism. On the average, 3.2% of

larvae died as a result of unknown causes (Table 4.1).

Cannibalism caused significantly greater mortality than

unknown causes (T=45.5; n=31; P<0.01), and exceeded unknown

causes in 22 of 31 cases. Numbers of deaths due to canni-

balism increased with increasing larval density (r=0.73;

P<0.0001). The percentage of larvae which died as a result

of cannibalism did not increase with increasing density

(r=0.15; P>0.40). Cannibalism was much more common between

first and second instar larvae (92% of total) than between

third instars. Third instar larvae often fed adjacent to

one another, and even contacted one another without

aggressive behavior or cannibalism.

Pupal weight decreased with increasing density for lar-

vae emerging from leaves on the seventh day after oviposi-

tion (day 7) (r=-0.15; P<0.0001), but larval density did not

affect weight of those emerging on day 8 (r=-0.04; P<0.50)

(Table 4.2). A scatterplot of the data from day 7 suggested

that the significance of the regression was due to pupal

weights of those from the three leaves with highest larval

densities (0.99 to 1.44 third instar larvae/cm2). When

pupal weights from the three leaves with highest densities

were excluded from the analysis, the regression was no

longer significant (r=-0.03; P>0.18). Inclusion of pupal

weights from the leaf with 0.99 larvae/cm2 with the lower








Table 4.2. Influence of L. sativae larval density on
weight and percentage survival of pupae.

x (S.E.M.) Pupal Biomass (Ag)

Density2 % Pupal
(larvae/cm2) n Day 7b n Day 8 Survivalc


0.21
0.21
0.27
0.29
0.36
0.39
0.42
0.43
0.43
0.45
0.49
0.49
0.52
0.54
0.54
0.56
0.57
0.57
0.60
0.60
0.62
0.65
0.66
0.68
0.72
0.72
0.73
0.79
0.99
1.15
1.44


439
457
471
424
440
450
432
452
463
400
389
436
491
426
460
421
412
459
406
427
454
443
422
413
427
432
472
438
395
419
398


(24)
(17)
(21)
(12)
(20)
(12)
(14)
(13)
(13)
(21)
(16)
(12)
(11)
(16)
(15)
(14)
(14)
(14)
(14)
(15)
(14)
(15)
(16)
(14)
(12)
(14)
(14)
(15)
(16)
( 9)
( 8)


5
3
5
9
4
5

5
5
6
11
12
2
10
2
6
10
6
18
18
10
16
13
8
12
7
9
5
11
21
17


514
453
550
464
450
534

450
466
520
449
425
515
422
460
483
452
423
536
504
475
464
524
452
472
503
477
482
458
486
452


(37)
(52)
(33)
(18)
(74)
(42)

(70)
(40)
(49)
(38)
(16)
(45)
(29)
(70)
(48)
(16)
(80)
(24)
(23)
(33)
(17)
(27)
(30)
(26)
(35)
(27)
(24)
(42)
(19)
(29)


82.6
87.5
81.5
80.0
70.8
94.4
66.7
96.9
91.7
81.5
85.0
87.0
90.0
88.6
91.5
91.7
81.6
95.2
79.6
80.8
90.9
96.3
83.0
81.8
87.5
90.7
84.8
90.6
86.3
95.4
87.5


881 434 ( 3)e 271 475 ( 6)

a Density of third instar larvae.
b Day after L. sativae oviposition on which larval
emergence from leaves occurred. Pupal weight
decreased with increasing larval density on day 7 (r=
-0.15; P<0.0001), but density did not affect pupal
weight of those emerging on day 8. (r=-0.04; P>0.50).
c % Pupal survival = (L. sativae adults/total puparia)
x 100%.
d No larvae emerged from this leaf after day 7.
e Mean weight of puparia resulting from larvae that
emerged on day 7 was significantly less than that
from larvae emerging on day 8. (LSD test; P<0.01).








densities resulted in a significant regression (r=-0.08;

P<0.009). Larvae emerging on day 7 from the three leaves

with the highest densities weighed significantly less than

those at lower densities (F=23.2; df=1,879; P<0.0001), while

pupal weights between these density groups were not differ-

ent for those emerging on day 8 (F=0.3; df=1,269; P>0.58)

(Table 4.3).



Table 4.3. Comparison of pupal weights for L. sativae
larvae reared at high and low densities.

Third Instar Day 7a Day 8
Density 2
(larvae/cm2) n x (S.E.M.) (gg) n x (S.E.M.) (Mg)

0.21-0.79 745 439 (2.9)a 222 476 (0.6)a
0.99-1.44 136 404 (6.0)b 49 468 (1.6)a

a Day after L. sativae oviposition on which larval emergence
from leaves occurred.



Pupal weights of larvae which had emerged on days 7, 8,

and 9 were significantly different (F=15.8; df=2,40;

P<0.005). Larvae emerging on day 7 weighed significantly

less than on those emerging on day 8 (P<0.01) (Table 4.2).

The eighteen larvae that emerged on day 9 weighed signifi-

cantly less than those emerging on either day 7 or day 8

(P<0.01).

Pupal survival was not affected adversely by increasing

third instar larval density (r=0.27; P>0.14). Pupal

survival was also not different between those larvae which

emerged on day 7 and day 8 (87.0 vs. 88.6%, respectively).








Overall, ca. 75% of the total number of larvae emerged

on the seventh day after oviposition, ca. 23% on day 8, and

2% emerged on day 9. Development times of larvae were not

affected by the density of third instar larvae within a leaf

over the range of densities tested (r=0.22; P>0.12).

Average leaf area mined by second and third instar L.

sativae larvae was 0.720.08 cm2 (n=25) (95% confidence

interval). This is approximately equal to the 0.630.16 cm2

mined by L. trifolii on lima bean primary leaves (J.T. Trum-

ble, personal communication, 1988). That L. trifolii on

chrysanthemum mined an average of 1.72 cm2 highlights the

impact of the host plant on leaf area mined (Parrella &

Bethke 1988).


Discussion
Cannibalism, a type of interference competition,

occurred primarily between first instar and second instar

larvae resulting in a maximum of 11.3% larval mortality.

The occurrence of cannibalism in the early instars was unex-

pected, because a large percentage of the total leaf area

mined is mined by the third instar and many more encounters

between larvae would be expected during this instar.

The extent to which cannibalism occurs in a leaf

appears to be affected by egg dispersion and restriction of

larval movement by leaf veins. Since egg dispersion was not

uniform, and movement of first and second instar larvae was

largely confined by leaf veins, cannibalism occurred more

frequently than would otherwise be expected during these








instars. The lack of aggressive behavior among third instar

larvae, along with their ability to move from place to place

within the leaf, resulted in little cannibalism occurring

during this instar. Quiring & McNeil (1984a) similarly

reported a lack of aggressive behavior for third instar

alfalfa blotch leafminers (Aqromvza frontella (Rondani)).

The significant negative correlation between pupal

weight and density among larvae emerging from leaves on day

7 is evidence that more exploitative competition occurred at

higher densities. The low correlation coefficient (r=-0.15)

indicates that pupal weight was not strongly affected by

density over the range of densities tested. Low values for

mean pupal weight at low density, e.g., 399 Ag at 0.49

larvae/cm2 (Table 4.2), suggest that competition can occur

in leaves with low larval density. Photographs of such

leaves revealed that heavy mining occurred between certain

leaf veins with little or no mining occurring in other areas.

Competition may have occurred among larvae emerging on

day 7, because the amount of available resource per actively

feeding larva may be lowest in the hours just prior to lar-

val emergence on day 7. Although third instars were

restricted less than earlier instars, leaf veins also

restricted their movement to some extent. By the time a

larva has reached the third instar, it has a limited time

period to feed before emerging on day 7 (ca. 24h at 250C,

see Chapter III). After ca. 75% of the total number of lar-

vae emerged on day 7, the total amount of remaining resource








was greatly diminished, but the amount available per larva

was apparently unaffected by larval density in this experi-

ment. That larvae which emerged on day 8 weighed signifi-

cantly more than those that emerged on day 7 was likely due

to the longer feeding period. However, at larval densities

higher than those observed in this experiment, little, if

any, additional leaf mesophyll tissue may remain for those

emerging on day 8.

Interference competition and other larval mortality did

reduce larval populations, but exploitative competition

still occurred among third instar larvae resulting in

reduced pupal weights at the highest densities studied.

Quiring and McNeil (1984b) also found both interference and

exploitative competition occurring among alfalfa blotch

leafminers. Oatman (1960), however, studying L. pictella

Thomson (probably _L. sativae, see Spencer 1981b) on bush

lima bean, reported no reductions in pupal weight with

increasing density, unless density was high enough to cause

leaf abscission.


Conclusions

Exploitative competition may occur even at low larval

densities if eggs are aggregated an a particular section of

a leaf. The available data indicate that exploitative com-

petition can be minimized by maintaining larval densities at

or below 0.8 third instar larvae/cm2. This density, below

which exploitative competition is minimized, however, is not






56

precisely defined, because of a scarcity of data points at

high densities.

If densities below 0.8 third instar larvae/cm2 are

used, even small lima bean primary leaves (ca. 55 cm2) can

support development of more L. sativae larvae than Ovius

dissitus was able to parasitize in one day (Petitt, unpub-

lished data, 1988). Thus, in oviposition studies that fol-

low (Chapter V), parasitoids can be provided adequate num-

bers of hosts on a single lima bean primary leaf.













CHAPTER V
EFFECTS OF PARASITIZATION OF Liriomvza sativae INSTARS ON
NUMBER AND CHARACTERISTICS OF Opius dissitus PROGENY


Introduction

The stage of host development and/or changes in the

host plant or environment associated with a particular stage

can affect the host selection process including host loca-

tion (Prince 1976, van Alphen & Drijver 1982, Glas & Vet

1983), and host acceptance (Hendrikse et al. 1980, van

Alphen & Drijver 1982, Liu et al. 1984). (See Doutt 1964

and Vinson 1976 for reviews of host selection.) The stage

of the host parasitized also affects suitability for the

parasitoid and host regulation. (See Vinson & Iwantsch

1980a, 1980b, Beckage 1985, and Lawrence 1986 for reviews.)

Parasitization of different host instars may influence

characteristics of adult parasitoid progeny that may be

related to fitness. Host size or host instar exposed to

parasitoids can affect the number of parasitoid progeny pro-

duced (Hendrikse et al. 1980, van Alphen & Drijver 1982),

and progeny sex ratio (Avilla & Albajes 1984, van den Assem

1984). Parasitoid size, which also may be affected by the

host instar parasitized, has been positively correlated with

egg production (Iwata 1966, Sandlan 1979, Lawrence 1981, and

Opp & Luck 1986), and/or the ability to obtain hosts (Boldt

1974, van den Assem 1976, Lawrence 1981). The number of








eggs in the ovaries of newly closed parasitoid females may

also be influenced by the host instar parasitized (Liu

1985).

The objective of these studies was to determine whether

parasitization of a particular L. sativae instar by 0.

dissitus would affect numbers, sex ratio, or size of

progeny, or numbers of eggs in progeny, when near-equal host

densities were provided.


Materials and Methods

Plant Culture and Host Development. Bush lima beans

were planted and maintained as described in Chapter II. On

three consecutive days, five pots of lima beans 10-14d old

were placed in a colony of adult L. sativae to allow ovipo-

sition for 6h on primary leaves. Plants were held in envi-

ronmental chambers for 3.8, 4.8, and 5.80.1d at average

temperatures of 23.6, 25.6, 25.60.20C to allow larvae to

develop to first, second, or third instars, respectively

(see Appendix B, Applications of Larval Development Model,

Example #1). Samples of ca. 20 host larvae were removed

from leaves at the beginning and end of the oviposition

interval to determine the percentage of larvae in each

instar. The lengths of mouth hooks and the cephalopharyn-

geal skeleton were used for instar determination (see Chap-

ter II). Exploitative competition between host larvae was

minimized by using host densities less than 0.8 larvae/cm2,

unless noted otherwise (see Chapter IV).








Numbers of progeny from various host instars. Equal

densities of each host instar were exposed to parasitoids to

compare the number of F1 progeny produced. It was diffi-

cult, however, to obtain equal host densities because the

number of ovipositions on leaves could be controlled only

crudely by adjusting the number of adult flies in the cage.

The number of L. sativae larvae in each leaf was counted

prior to exposure to parasitoids and leaves with unusually

high or low densities were not used. Larval density could

only be estimated at this time because passage through the

leaf area meter (Li-CorR Model 3000, Lincoln, Nebraska)

would have caused larval mortality. As a result many leaves

were included in the experiment which eventually were elimi-

nated prior to analysis, because densities of a particular

host instar were not in the same range as those available

for each of the other host instars.

0. dissitus adult females were collected from a green-

house colony and held overnight in a circular plastic dish

(see description of oviposition arena below) with an equal

number of males as described above. A 10% sucrose solution

was provided on small sponges as food. The food source was

included because preliminary experiments indicated that con-

siderable adult mortality occurred when no food or hosts

were provided. The following morning, one female was

released at random into each oviposition arena. Each arena

contained a single primary lima bean leaf with host larvae

of a particular instar. Oviposition arenas were circular








plastic dishes 17cm in diameter by 5cm in height (Shamrock

Industries, Minneapolis, MN) with an organdy lid. A water

reservoir (Syndicate Sales Inc., Kokomo, IN) holding the

petiole of an excised lima bean leaf projected through the

side of the arena. Arenas were randomly assigned positions

in a controlled environment chamber at 2510C and 7010%

relative humidity. Host larvae were exposed to parasitoids

for a 6h period. Leaves were kept in the arena until all

host larvae had emerged and pupariated. Leaf area was mea-

sured upon removal. Arenas remained in the chambers under

the same conditions until all adult hosts and parasitoids

had been collected (ca. 18d after parasitoid oviposition).

The total number of puparia and those from which no adult

eclosion had occurred were counted. Four replicate experi-

ments were conducted.

Body length of male and female progeny was measured

from the frons to the tip of the abdomen excluding the

ovipositor in females. Measurements were made to the nearest

0.0325mm under a dissecting microscope at 180x total magni-

fication. Progeny from only two replicate experiments were

measured, because inadequate numbers of progeny of each sex

were available from first instar hosts in the other two

replicate experiments.

Effect of Host Instar Parasitized on Number of Eggs in

Ovaries of Parasitoid Progeny. Lima bean plants with pri-

mary leaves containing either larvae in the first, second,

or third instars were placed in separate cages (ca. 0.07m3).








One-day-old Q. dissitus adults were randomly divided into 3

groups (1:1 sex ratio) and released into the cages to

oviposit on a particular host instar for 6-8h. Prior to

each experiment, parental females were kept overnight with

males as described above. Primary leaves were excised ca.

12h before larval emergence began, placed in water reser-

voirs filled with a modified Hoagland solution and suspended

over polypropylene funnels. Larvae emerging from leaves

were collected and allowed to pupate in 30ml diet cups

placed at the base of the funnel. From infestation through

adult emergence all diet cups were kept in the same environ-

mental chamber maintained at 2510C and 6515% relative

humidity with 14h photoperiod beginning 0700 hours. Adult

female parasitoid progeny were collected within 4h of eclo-

sion and preserved in a 70% ethanol:water mixture until dis-

section was conducted to determine the number of chorionated

eggs in their ovaries. Ten female progeny each resulting

from parasitization of first, second, and third instar hosts

were dissected in each of 4 replicate experiments. To

determine whether body length was correlated with the number

of chorionated eggs in newly closed females, 16 females

were measured and then dissected.

Data Analysis. Progeny from females which produced

only males were not included in the calculation of sex

ratio. Because 0. dissitus exhibits haplodiploid reproduc-

tion, the production of only male progeny may have resulted

from the female being unmated. Progeny sex ratio was arcsin








square root transformed before analysis as recommended by

Steel and Torrie (1960) for percentages which cover a wide

range of values.

A completely randomized design was used in each repli-

cate experiment. Data from the replicate experiments

designed to determine numbers and size of progeny resulting

from exposure of equal densities of each host instar were

first analyzed using one-way analysis of variance with homo-

geneity of variance tested using Bartlett's Box F-test

(Oneway Procedure; SPSSX 1986). An analysis of variance was

then performed with data from the replicate experiments com-

bined (Cochran & Cox 1950). Bartlett's method was used to

test for homogeneity of experimental error variances from

the four experiments (Sokal and Rohlf 1981). The experiment

by instar interaction was tested using the pooled experimen-

tal error. Since heterogeneity operates so that tabular F

produces too many significant results, an F-test of the

experiment by instar interaction that is not significant is

still valid in the presence of heterogeneity (Cochran & Cox

1950). The experiment by instar interaction was never

significant in these experiments (P>0.05). Consequently,

the interaction mean square was used to test instar effects.

An additional assumption required for the F-test of

treatments against experiment by instar interactions is that

interaction variances are homogeneous (Cochran & Cox 1950).

Because this assumption cannot be tested, the exact

probability level of the resulting F-value will not be








known. However, the upper limit to the significance of F

was determined as recommended by Cochran & Cox (1950). With

four experiments in the combined analysis, the upper limit

to the significance of F is distributed approximately with 1

and 3 degrees of freedom (Cochran & Cox 1950).

The "protected" LSD test was used to separate means

because it guards against a high experiment-wise error rate

(Jones 1984). Data from replicate experiments used to

determine whether parasitization of different host instars

affected the number of eggs in progeny also were analyzed in

a combined analysis.

Voucher Specimens. Voucher specimens of L. sativae

were collected and submitted as described in Chapter II.

R. A. Wharton (Texas A&M Univ.) identified specimens of 0.

dissitus and voucher specimens were deposited in the Florida

State Collection of Arthropods in Gainesville, Florida.


Results

Numbers and size of progeny from various host instars.

The method described to obtain first, second, and third

instar L. sativae larvae simultaneously for use in the

experiments was quite reliable. On the average, 100, 95.0,

and 83.3% of larvae expected to be in the first, second, and

third instar were confirmed to be in that instar during the

6h experiments (see Appendix D, Table D.1 for instar dis-

tribution in each replicate experiment). The lowest per-

centages of larvae in the first, second, and third instars








in any of the replicate experiments were 100, 87.5, and

74.6%, respectively.

The density of host larvae in the four replicate exper-

iments averaged from 0.27 to 0.55 larvae/cm2. The average

density of first, second, and third instar hosts in any sin-

gle experiment varied by not more than 0.08 larva/cm2. The

average number of host larvae provided to a parasitoid in

the four experiments was 29.8 (S.E.M.=1.20). While the min-

imum number of host larvae provided to a parasitoid was 14,

87% of the parasitoids were provided 20 or more larvae.

Numbers of 0. dissitus progeny produced during the 6h

oviposition interval were significantly affected by the host

instar which was exposed to the parasitoid (F=11.40; df=2,6;

P<0.01) (Table 5.1). The 0.01 probability level assumes

homogeneity of interaction variances. The modified signifi-

cance level suggested by Cochran & Cox (1950), which is

valid in the presence of heterogeneity, showed significance

below the 0.05 level (F=11.40; df=1,3; P<0.05). Signifi-

cantly more parasitoid progeny resulted from exposure of

second and third instar hosts. The sex ratio of progeny was

slightly female biased, but did not vary with host instar

parasitized (F=0.25; df=2,6; P>>0.10).

Female size was affected by the host instar parasitized

(F=110.56; df=2,2; P<0.01) (Table 5.2). Smaller females

resulted from parasitization of third instar hosts (P<0.05;

"protected" LSD test). Male size, however was not affected




















Table 5.1. Numbers of 0. dissitus progeny produced when
parental females exposed to near equal
densities of different host instars.

Progeny
Host x (S.E.M.) Sex Ratio
Instar n Progeny/Femalea (% Female)


1 31 5.9 (1.16) a 58.8

2 26 9.3 (1.32) b 54.3b

3 26 11.6 (1.63) b 56.3


a Combined analysis of 4 replicate experiments. Exper-
iment by instar interaction was not significant.
Interaction mean square was used to test instar
effect (F=11.40; df=2,6; P<0.01; df=1,3; P<0.05).
Means followed by the same letter are not signifi-
cantly different (P=0.05; "Protected" LSD test).
Females producing only male progeny were excluded
from the calculation and analysis of sex ratio. Com-
bined analysis of 4 replicate experiments. Experi-
ment by instar interaction was not significant.
Interaction mean square was used to test instar
effect (F=0.25; df=2,6; P>>0.10)




















Table 5.2. Size of 0. dissitus progeny resulting from
parasitization of various larval instars of
L. sativae.


Host x body length (S.E.M.) (mm)a,b
Instar
Parasitized
n Females n Males

1 42 1.560 (0.0163) a 47 1.552 (0.0191) a

2 63 1.555 (0.0143) a 64 1.585 (0.0152) a

3 75 1.495 (0.0147) b 48 1.555 (0.0151) a

a Length measured from frons to tip of abdomen (excluding
ovipositor in females). Combined analysis from 2
replicate experiments. Experiment by instar interaction
was not significant in either case. Interaction mean
square was used to test instar effect (Females,
F=110.56; df=2,2; P<0.01; Males, F=0.60; df=2,2;
P>0.50).
b Means followed by the same letter are not significantly
different. (P=0.05; "Protected" LSD test).








by the host instar parasitized (F=0.60; df=2,2; P>0.50).

Males and females of this species are very similar in size.

Effect of host instar parasitized on number of eccs in

ovaries of parasitoid progeny. On the average, 100, 94.0,

and 93.1% of larvae expected to be in the first, second, and

third instars were confirmed to be in the correct instar

during the 6h experiments (see Appendix D, Table D.2 for

instar distribution in each replicate experiment). The low-

est percentages of larvae confirmed to be in the first, sec-

ond, and third instars in any of the experiments were 100,

87.5, and 90.1%, respectively. The highest host density in

any cage exposed to parasitoids was 0.89 third instar

larvae/cm2.

Each female has four ovarioles and the reproductive

system resembled that of Biosteres (=Diachasma) tryoni (Pem-

berton & Willard 1918). Chorionated eggs appeared a shiny

white when light reflected from them. Female progeny

contained from 0 to 24 chorionated eggs with an average of

ca. 12 eggs. The mean number of chorionated eggs in female

progeny resulting from parasitization of first, second, and

third instar hosts were not significantly different (F=1.56;

df=2,6; P>0.10) (Table 5.3). Body length was not correlated

to the number of chorionated eggs in newly emerged females

(r=0.17; P>0.26).


Discussion
Previous studies with larval-pupal parasitoids of

dipterans suggest that larger numbers of progeny resulted

















Table 5.3. Mean number of chorionated eggs in female
progeny of 0. dissitus resulting from
parasitization of different instars of L.
sativae.


Host Instar x Eggs/Female
Parasitized n Progeny (S.E.M.)a

1 40 13.5 (0.53) a

2 40 12.5 (0.71) a

3 40 11.3 (0.67) a

a Combined analysis from 4 replicate experiments.
Experiment by instar interaction was not significant.
Interaction mean square was used to test instar
effect (F=1.56; df=2,6; P>0.10).








from exposure of later instar hosts because these hosts are

more easily located. Hendrikse et al. (1980) found that, on

the average, Dacnusa sibirica Telenga and Opius pallipes

Wesmael required ca. 2.5 to 3.4 times longer to locate small

Liriomyza bryoniae Kalt. larvae than medium or large larvae.

This may be due to the smaller magnitude of vibrations or

less frequent movements of small larvae, as observed in

Drosophilia (Prince 1976, van Alphen & Drijver 1982) or pos-

sibly a lower concentration of kairomone associated with

smaller hosts. Even if vision alone were used to locate

hosts, the large leaf mine of the older hosts would be more

often encountered by a searching parasitoid than a small

mine. Hendrikse et al. (1980) and Petitt (1984) have

observed that when a parasitoid encounters a mine, it very

frequently locates the host itself shortly thereafter.

It is possible that for mass-rearing of 0. dissitus to

be economical, host densities would have to be increased

above the highest densities experienced in these oviposition

studies. In that case, I suspect that the relative numbers

of progeny produced from exposure to different instars would

likely be similar to the results found here, because the

same host location mechanisms would likely be used. How-

ever, because it is possible that the effectiveness of host

location mechanisms may change with host density, an experi-

ment at mass-rearing density may be warranted. It is

unlikely that first instar larvae would be optimal for rear-

ing at high host densities. Provided that parasitized hosts






70

behave similarly to unparasitized ones, parasitoid mortality

would increase with density due to cannibalism between early

instar hosts (see Chapter IV).

Although differential parasitoid mortality could occur

due to cannibalism between first instar host larvae, it can-

not explain the large differences in numbers of progeny that

resulted from parasitization of different host instars in

this study. Previous experiments indicate that less than 5%

larval mortality should result from cannibalism in the den-

sity range studied (see Chapter IV).

Few studies on the effect of host age or host instar on

progeny sex ratio have been reported in larval-pupal para-

sitoids of dipterous hosts or in other parasitoids which

attack hosts which continue to feed, grow, and develop after

parasitization. In one such study, Avilla and Albajes

(1984) found that when single Opius concolor Szepl. were

provided older hosts the sex ratio was more female-biased.

In this study, however, Q. dissitus sex ratio was not

affected by host instar parasitized.

In another larval-pupal parasitoid of a dipteran host,

progeny sex ratio became more male-biased with increases in

density of parasitoid females (Lawrence 1981). This sug-

gests that the progeny sex ratio of 0. dissitus may be more

male-biased in a mass-rearing situation than in these exper-

iments where females were held in separate containers. How-

ever, in several years of laboratory rearing of 0Q. dissitus,








no strongly biased sex ratios have been observed (Petitt,

unpublished data, 1988).

Females resulting from parasitization of third instar

hosts were smaller in length than those which were para-

sitized in earlier instars. In studies with other para-

sitoids, however, parasitization of first instar hosts

resulted in smaller females (Lawrence et al. 1976, Liu

1985). In 0. dissitus, it is possible that prolongation of

the first stadium of the parasitoid in hosts parasitized in

the first or second instar (Chapter VI) may affect adult

body size by increasing the time available for feeding by

this instar. It is also possible that hemolymph properties,

such as concentration of proteins and free amino acids,

change drastically during these larval stadia, affecting

nutritional suitability of the host for the parasitoid

(Corbet 1968).

In other parasitoid species, positive relationships

have been reported between female body size and egg produc-

tion (Iwata 1966, Sandlan 1979, Lawrence 1981) and between

size and the ability of females to obtain hosts (Boldt 1974,

van den Assem 1976, Lawrence 1981). The impact of the

observed size difference on characteristics which may affect

fitness of 0. dissitus, however, is not known.


Conclusions
To obtain largest numbers of progeny, Q. dissitus

adults should be provided second or third instar hosts in

which to oviposit. Even if desirable, it would be very








difficult to provide exclusively second instar host larvae

to parasitoids, particularly if they are reared under green-

house conditions where only crude environmental control is

possible. (Total duration of second instar is ca. l.ld at

250C. See Chapter III.) For this reason, either second

and/or third instar hosts should be provided to parasitoids

for oviposition.

Female parasitoid size was affected by host instar par-

asitized. It is not known whether the slightly smaller size

of females reared from hosts parasitized as third instars is

of biological significance. Better methods of determining

parasitoid quality are needed, but are extremely difficult

to develop. It would be desirable to find measurable char-

acteristics which relate to the realized fitness of the par-

asitoid in environments where they are released for leaf-

miner control.













CHAPTER VI
EFFECTS OF TEMPERATURE AND HOST INSTAR
ON DEVELOPMENT TIME OF Opius dissitus


Introduction

Insect development times are greatly influenced by tem-

perature. Determining how development times change with

temperature is a valuable step toward understanding certain

aspects of population dynamics of both pests and their natu-

ral enemies. For example, the course of ecological interac-

tions critical to successful biological control may be

affected by thermal requirements of the species involved

(Campbell et al. 1974, Frazer & Gilbert 1976, van Lenteren &

Hulspas-Jordaan 1983). In addition to its usefulness in

studies of population dynamics, knowledge of the effects of

temperature on development of the pest and its natural enemy

are required to rear the natural enemy for laboratory evalu-

ation or to properly time releases of the beneficial insect

into the field. The effects of temperature on the develop-

ment of L. sativae larval instars were previously reported

(Chapter III) and studies of 0. dissitus development are

reported here.

Because all three host instars can be parasitized suc-

cessfully by 0. dissitus (Chapter V), and host instar is

known to affect development of endoparasitic Hymenoptera

(Corbet 1968, Vinson & Barras 1970, Lawrence et al. 1976,








Sato 1980), the effect of host instar parasitized on devel-

opment of 0. dissitus was studied. In the course of these

experiments, it was determined that parasitization of second

and third instar hosts resulted in very similar parasitoid

development times at 250C. Numbers of parasitoid progeny

produced on these host instars were not significantly dif-

ferent (Chapter V). Consequently, the remaining temperature

studies reported here were conducted using a mixture of

second and third instars.

The primary objective of the temperature studies was to

develop a linear degree-day model for egg-to-adult develop-

ment of 0. dissitus. Accurate predictions of development

time should be possible at the intermediate temperatures

used in rearing (Wagner 1984). The secondary objective was

to determine the feasibility of separation of unparasitized

hosts from the 0. dissitus destined for release. Parasi-

toid-induced changes reported in many host insects (see Vin-

son & Iwantsch 1980, Beckage 1985, and Lawrence 1986 for re-

views), suggested that timing of larval emergence may be af-

fected by parasitization. Alternatively, adult eclosion of

the host and parasitoid species may be temporally separated.


Materials and Methods

Effect of host instar on develoDment. Bush lima beans

were planted and maintained as described in Chapter II. On

three consecutive days, five pots of lima beans 10-14d old

were placed in a 1.9 x 1.1 x 0.6m cage containing a colony

of adult L. sativae to allow oviposition for 6h on primary








leaves. Plants were held in environmental chambers for 3.8,

4.8, and 5.80.1d at average temperatures of 23.6, 25.6,

25.60.20C to allow larvae to develop to first, second, and

third instars, respectively. Plants were placed in cages in

an environmental chamber with parasitoids for 6h beginning

0900 hours. Samples of ca. 20 host larvae were removed from

leaves at the beginning and end of the oviposition interval

to determine the percentage of larvae in each instar (see

Appendix E). Larval mouth hooks and/or the cephalopharyn-

geal skeleton were measured for instar determination (see

Chapter II). Based on development time data (Chapter III),

primary leaves were excised ca. 12h before larval emergence

began, placed in water reservoirs (Syndicate Sales Inc.,

Kokomo, IN) filled with a modified Hoagland solution (see

Appendix A) and suspended over polypropylene funnels. Lar-

vae emerging from leaves were collected and allowed to

pupate in 30 ml diet cups placed at the base of the funnel.

From infestation through adult emergence all diet cups were

kept in the same chamber maintained at 2510C and 6515%

relative humidity with a 14h photoperiod beginning 0700

hours. Adult emergence of L. sativae and 0. dissitus was

observed at 0700, 1100, and 1500 hours and parasitoid sexes

were differentiated by the presence of the ovipositor in

females.

To determine which parasitoid stage or stages were pro-

longed when oviposition occurred on first instar hosts, host

development was scheduled to obtain first and third instar








larvae in lima bean primary leaves as described above. One-

to three-day-old 0. dissitus oviposited on first and third

instar hosts for 6h. Plants and puparia were maintained

thereafter in an environmental chamber at 2510C. Dissec-

tions of at least 10 larvae or pupae were made daily from

each group for 5 days after oviposition.

Effects of temperature on development. Plant culture

was the same as described in Chapter I. L. sativae ovi-

posited on 15 lima bean plants for 6h in a greenhouse cage.

Plants were held in environmental chambers at 251lC for ca.

5.8d until second and/or third instar larvae were present.

Plants were placed in a large greenhouse cage where para-

sitoid oviposition occurred for 6h. Five plants each were

returned to environmental chambers at constant temperatures

of 15, 20, 25, 30 or 350C and a 14h photophase beginning

0700 hours. A CR7 datalogger (Campbell Scientific, Logan,

UT) monitored temperatures every minute and maintained

running averages from which degree-days were calculated as

described below in the Data Analysis section. Host larvae

began to emerge from leaves the day after oviposition and

were collected into diet cups every 2h from 0800 to 1600

hours daily. Cups were checked at the start of the photo-

phase (0700 hours) to see if any emergence had occurred

overnight. Adult emergence was observed daily at the same

2h intervals and parasitoid sex was determined. The exact

time of adult emergence of the 0. dissitus collected at the

beginning of the photophase was not known, but it was








assumed to be 0600 hours. The experiment was conducted

twice at each temperature, except a third experiment was

conducted at 150C and 300C where dissimilar outcomes

occurred in the first two experiments.

In one experiment at 20, 25, 30 and 350C, the time

required from parasitoid oviposition to larval emergence

from leaves was compared for parasitized and unparasitized

larvae to determine whether there were any parasite-induced

changes in this behavior. In two experiments at 250C,

records of adult eclosion of both the host and parasitoid

were recorded.

Data Analysis. Completely randomized designs were used

for the three replicate experiments testing the effect of

host instar on parasitoid development time. Data were first

analyzed using a one-way analysis of variance with homogene-

ity of variance tested using Bartlett's Box F-test (Oneway

Procedure; SPSSX 1986). An analysis of variance was then

performed with the data from the replicate experiments com-

bined (Cochran & Cox 1950). The experiment by instar inter-

action was tested using the pooled experimental error. The

experiment by instar interaction was never significant

(P>0.05). Consequently, the interaction mean square was

used to test the instar effects. Bartlett's method was used

to test for homogeneity of experimental error variances from

the three experiments (Sokal & Rohlf 1981). The "protected"

LSD test was used to separate means because it guards

against a high experiment-wise error rate (Jones 1984).








The analysis of development times of males and females

within an instar were also analyzed using a combined analy-

sis. The experiment by sex interaction was not significant

(P>0.05), consequently, the interaction mean square was used

to test the effect of sex. Because only two means were com-

pared, no mean separation test was required.

Linear regression used to relate development rates to

temperature was conducted using the SPSSX Scattergram proce-

dure (SPSSx 1986). The lower developmental threshold (LDT)

was calculated as the x intercept of the regression line.

The number of degree-days (DD) required for egg-to-adult de-

velopment of 0. dissitus was calculated for each individual

as follows:

DD = (T LDT) (Dt),

where T was the average temperature in the environmental

chamber, and where Dt was the average time in days required

for 0. dissitus to develop from egg to adult. The mean and

standard error in number of degree-days required for 0.

dissitus to develop at constant temperatures up to 29.80C,

inclusive, were calculated.

Voucher Specimens. Voucher specimens of both species

were collected and submitted as described in Chapter V.


Results

Effects of host instar on development. On the average,

100, 96.2, and 92.4% of L. sativae larvae expected to be in

the first, second, and third instars were confirmed to be in






79

the correct instar during the 6h experiments. (See Appendix

E for instar distribution in each replicate experiment.)

Host instar significantly affected egg-to-adult devel-

opment time of 0. dissitus (Table 6.1). Development time

was shortest for the parasitoid on third instar hosts, ca.

0.3d longer, on the average, for those on second instar

hosts and ca. 2d longer on first instar hosts.

Daily dissections of host larvae revealed that the du-

ration of the parasitoid's first instar larval stadium was

prolonged when females oviposited in first instar hosts

(Fig. 6.1). Regardless of the host instar parasitized, how-

ever, 73% or more of parasitoids underwent the first larval

molt between 24 and 48h after host pupariation. The remain-

der underwent the first larval molt beyond 48h after host

pupariation.

On the average, male 0. dissitus in these experiments

always developed more quickly than females (Table 6.1). The

difference in development times between the sexes on second

instars was not significant because variability in the

extent to which males developed faster than females in the

three experiments led to a large (but not significant)

experiment by sex interaction mean square. When this mean

square was used to test the effect of sex, it was not sig-

nificant (F=14.95; df=1,2; P<0.10).

Effect of temperature on development. Egg-to-adult

development time for 0. dissitus decreased from ca. 50d at

150C to ca. 11d at 340C (Table 6.2). Development time in














Table 6.1.


Egg-to-adult development times for 0.
dissitus resulting from parasitization of
different instars of L. sativae at 250C.


Instar Analysisa Sex Analysisb

n x Dev. Time (S.E.M.) Sex n x Dev. Time (S.E.M.)
(days) (days)


1 181 14.60 (0.059) a M 88 14.42 (0.081) a
F 93 14.77 (0.081) b

2 189 12.82 (0.052) b M 85 12.50 (0.063) a
F 104 13.08 (0.070) a

3 179 12.51 (0.056) c M 103 12.26 (0.067) a
F 76 12.85 (0.080) b

a Combined analysis from 3 replicate experiments.
Experiment by instar interaction was not significant.
Interaction mean square was used to test instar effect
(F=399.25; df=2,4; P<0.005). Means followed by a
different letter are significantly different (P<0.01;
protected LSD test).
b Male vs. female development times compared within an
instar across experiments. Experiment by sex
interaction was in no case significant. Interaction
mean square was used to test effect of sex (first
instar, F=30.70, df=1,2, P<0.05; second instar,
F=14.95, df=1,2 P<0.10; third instar, F=62.18, df=1,2,
P<0.025). Means within an instar followed by a
different letter are significantly different by F-test
(P=0.05).




















Host Parasitized as:
I 1st Instor El 3rd Instar
120 -


100


80-


0'


IL


t, j 4
Time After Parasitoid Oviposition (Days)


Fig. 6.1. Percentage of Q. dissitus in the first sta-
dium within L. sativae parasitized either
as first or third instar larvae. (Hosts
parasitized as third instars pupariated on
the first day after parasitoid oviposition,
while those parasitized as first instars
pupariated on the third day after para-
sitoid oviposition.)


I















I I


rO
N
I H



cO
C.,


0
c

0




U)
(11


0 0
30


0 W



-4 0)





-44-
P 44



0











>9 1
m



010














r q
o *P-
0) (0






* 01
S0

0
N







0)
0 3

U)
ag




I E-



0>C










E4e


Sin Co H N Hr
H Hd d H
n il rl < rH


w 0 0- N 01 w
N


H-4 M r- N
H N


H n N co






H-HrH
i-l i-l r


1nr H
in v


'D C. ofcmr-
H-4 e LU .I


'nO HCo n0
H


IV N
* *

0 co





0 0 0 0
~OUC



QIQ 0-1 r


4- ) 0
o .,
0 4.

Q 0
0 O

0
a)


OD


00
o W

r-4 *
S3 a -H




I.490 0




ac4 O
> .9 C (





a 90 400

r-4 41 41
.C .4. 4



V *a

I 4 + Oa

tO no g
r-4 3 tv C *
o o'04.' n


H a%4 o
> 041 0
(MU
() 0 r 0 ro




4o 0
0rl 4J ) (0
O4 0 04 4


0 0 w)i
a a c o
0 .o ( U
H> 4 04 4
t 00 0 H




1 go A

k 0 04 rO
0M 0 34
(0 A- 0 + (

Q U PO-PC
tT~r-l C
cr0P)(0Q0Q


O cO r(
C., ( (V4CO








one of three replicates at 150C was unlike the other two and

ranged from 35 to 120d with a mean of 74d. Occurrence of

adult eclosion over this long time span was similar to that

described for Opius fletcheri following diapause (Willard

1920), and, consequently, the point was not included in any

analysis. Development times decreased with increasing

temperature, but did not decrease further at temperatures

above 300C.

Development rate increased linearly with temperature up

to 300C, allowing application of a linear degree-day model

over this range. Dissimilar development times were observed

for the three replicate experiments near 300C. The observed

development time was ca. Id less at 29.80C than at 29.7 and

29.60C (Table 6.2). This likely was due, in large part, to

the differences in the development times on second and third

instar host larvae near 300C. In the experiment at 29.80C,

host larvae were predominantly third instars at the time of

parasitoid oviposition (day 6), as indicated by the fact

that 85% of them emerged from leaves on day 7. At 29.7 and

29.60C, however, only 15% and 26% of host larvae emerged on

day 7, respectively, indicating that they were predominantly

second instar larvae at the time of parasitoid oviposition.

From these data it appears that the small delay in develop-

ment (ca. 0.3d) observed when parasitoids oviposited on sec-

ond instar hosts at 250C (see Table 6.1) increased to a

delay of ca. Id at temperatures near 300C.








The regression equation relating egg-to-adult develop-

ment rates (y) to temperature (x) was y = 0.0052x 0.0558

(r = 0.99). It had an x intercept of 10.80C which was used

as the lower developmental threshold (LDT). The mean number

of degree-days required for egg-to-adult development was

quite consistent across temperatures, but the variability

was higher for the experiments conducted at ca. 150C (Table

6.2).

At ca. 200C or above, the egg-to-adult development time

of 0. dissitus was shorter than those of its host, L.

sativae (Fig. 6.2). (Data for L. sativae from Chapter III.)

For example, at 250C the development time of the parasitoid

was 4d shorter than its host. At ca. 150C, however, the

mean development time of 0. dissitus was very similar to

that of L. sativae (ca. 50d). The LDT determined for the

parasitoid (10.80C) was higher than the LDT for egg-to-adult

development of its host (10.30C, r=0.99) over the same

temperature range.

In these experiments, the parasitoid did not induce

changes which significantly affected the time of host emer-

gence from leaves (Table 6.3). On the average, 93% of host

larval emergence from leaves (both parasitized and unpara-

sitized) occurred during the first 9h of the photophase, and

89% (n = 2028) occurred between 0700 and noon.

Adult eclosion of L. sativae and 0. dissitus also

occurred primarily during the first 9h of the photophase

(Fig. 6.3). Eclosion of 98.4% of L. sativae and 90.3% of









































101 1 1 2I
15 20 25 30 35
Temperature (oC)


Fig. 6.2.


Development time of L. sativae and 0. dis-
situs in constant temperature experiments
from 15 to 350C.




















a)
at
N
4
*4.1


0)
0
4
(0
ca


4M >
0 )
0)

c)
a o
C)





4
01


*a >
01 4
40 i

0
*r-








0 4
-r
t



0




10
a)
wa)
QiN

U-)
01










EQ


0 0 0 0
C' C M C%




o c n o

N CV N







cf) H-4 N H
l OD m







(0 0 0 0


co Ln N %0
(n 0 H0 N




N N N CN










N H











0 0 0 0
o in o in
N N en


0









a)
0)


a)
06-


ta
S4)



4-1 (1






O
0%





l to
0)
G44

V





44



to -
0 4-




44
0U
0 3






4-l
0)

















I 0. di ss i tus (n-1o68)


20

10 1111
00-02 04-06 08-10 12-14 16-18 20-22
02-04 06-08 10-12 14-16 18-20 22-24
Time of Daq

Fig. 6.3. Frequency distribution for time of day of
adult eclosion for L. sativae and 0.
dissitus.


nh


0 L. sat i vae (n-387)


60 -

50 -








Q. dissitus occurred during these hours. The remaining 9.7%

of Q. dissitus adults closed between 1600 and 0700 hours

the next morning.

At 250C, eclosion of host adults began one or more days

prior to adult eclosion (Table 6.4). In experiment 1,

eclosion of both species occurred on days 11, 12, and 13,

whereas in experiment 2 overlap in adult eclosion occurred

only on day 11. The lower average temperature of ca. 0.50C

in experiment 1 may explain the longer development times.

In each experiment, on the day when maximum overlap in eclo-

sion of hosts and parasitoids occurred, 92% of the closing

parasitoids were males.


Discussion

The delay in molting to the second instar of ca. 2d for

those parasitoid larvae in hosts parasitized as first

instars accounted for the longer egg-to-adult development

time for this group. Synchronization of a parasitoid's molt

to the second instar with host metamorphosis has been

observed with several parasitoids of dipterous hosts (Pem-

berton & Willard 1918, Lathrop & Newton 1933, Lawrence

1982). Although the biochemical mechanism has not been

investigated with 0. dissitus, the parasitoid may be utiliz-

ing the ecdysteroid-mediated biochemical changes in the host

as a cue to the initiation of its own molt (Lawrence 1986).

Reports of shorter development times for male para-

sitoids are common (Doutt 1964, Lawrence et al. 1976) and

agree with observations on 0Q. dissitus. Because male














0





U
to
N

4-i.
0)




01





Va)
01 IV





Ow
> 0







01





o
.r-4




CM
9U0
0 V



W 0

*o
.0



41




H O
jQ)
0


-P




N r-H

C 0

0
.* 44
S0 4)

.0 4 II
0 0
o
4*














c 00
5






S *H r
0) m













WaO
0\0 -i
*4 0
m 0

(d0 04
v v


a% ON %D
H (n N


m 0 0


C% o 0 H- N m v fO
S- H H- H H


0o 0 ) 0 al N
-H I)












% u, 0 0 0 0
r- H











o 0 N wC MO
n

r.


* a
Nc'

WaY




0(0




x x
WCa





Wa)





Wa

Wa)

to Io



0) 0)




$4
OW
(0 XI








parasitoid wasps usually mate many times and females only

once (Matthews 1982), reduced development time may enable a

male to increase the number of virgin females he can fertil-

ize (King 1987).

Effects of temperature on development. Egg-to-adult

developmental studies at different temperatures were con-

ducted on a mixture of second and third instar larvae

because the results of host instar studies indicated that

only a small delay in development (0.3d) resulted from use

of second instar larvae. The assumption that host instar

would only very slightly affect development time at other

temperatures did not hold true. At 300C, parasitoids took

ca. ld longer, on the average, to develop on second instar

hosts than on third instar hosts. Slightly slower develop-

ment on second instar hosts possibly caused individuals to

miss "gates" or "allowed zones" for adult eclosion in this

light:dark regime (sensu Saunders 1982), thereby increasing

mean development time. Mackauer and Henkelman (1975) also

observed discontinuous adult eclosion in Aphidius smithii

Sharma and Subba Rao under alternating light:dark condi-

tions.

0. dissitus has shorter development times than L.

sativae at or above 200C. This is considered a positive

attribute of a natural enemy being evaluated for use in bio-

logical control (van Lenteren 1986). It suggests that, at

these temperatures, parasitoids may be able to respond

numerically to increases in the host population. At 150C,








however, the development times of these species were very

similar. Although no data were collected at lower tempera-

tures, the lower developmental threshold calculated by lin-

ear extrapolation for 0. dissitus (10.80C) was higher than

that calculated for L. sativae (10.30C). This suggests that

0. dissitus will develop more slowly than L. sativae at tem-

peratures near the LDT. However, the development of one or

the other of these species may respond non-linearly to tem-

perature in this range.

Campbell et al. (1974) compared the LDT's for several

aphids and their parasitoids and similarly found that the

LDT's for parasitoids were always higher than their hosts.

It was suggested that this might be a mechanism to ensure

that hosts become well established in a growing season

before parasitism begins (Campbell et al. 1974). Unlike the

results in this study, however, Campbell et al. (1974) found

that the aphid parasitoids always developed more slowly than

their hosts.

That no significant changes in timing of host larval

emergence occurred with parasitism is similar to observa-

tions with other larval-pupal parasitoids of dipterous hosts

(Lathrop & Newton 1933, Lawrence 1986). The apparent lack

of parasitoid-induced changes in these dipterous hosts could

be related to their nutritional or physiological adequacy to

support growth of parasitoid eggs and first instar larvae

(Lawrence 1986).