Citation
Biogeography of braconid parasitoids of the Caribbean fruit fly, Anastrepha suspensa (Loew) (Diptera: Tephritidae), in Florida

Material Information

Title:
Biogeography of braconid parasitoids of the Caribbean fruit fly, Anastrepha suspensa (Loew) (Diptera: Tephritidae), in Florida
Creator:
Eitam, Avraham, 1960-
Publication Date:
Language:
English
Physical Description:
viii, 183 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Cherries ( jstor )
Female animals ( jstor )
Fruit flies ( jstor )
Guavas ( jstor )
Larvae ( jstor )
Loquats ( jstor )
Low temperature ( jstor )
Parasite hosts ( jstor )
Parasitism ( jstor )
Parasitoids ( jstor )
Anastrepha -- Florida ( lcsh )
Braconidae -- Geographical distribution -- Florida ( lcsh )
Dissertations, Academic -- Entomology and Nematology -- UF ( lcsh )
Entomology and Nematology thesis, Ph. D ( lcsh )
Parasitoids -- Florida ( lcsh )
City of LaBelle ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 171-182).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Avraham Eitam.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029225613 ( ALEPH )
41186496 ( OCLC )

Downloads

This item has the following downloads:


Full Text










BIOGEOGRAPHY OF BRACONID PARASITOIDS OF THE CARIBBEAN FRUIT FLY, ANASTREPHA SUSPENSA (LOEW) (DIPTERA: TEPHRITIDAE), IN FLORIDA












By

AVRAHAM EITAM


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


1998













ACKNOWLEDGEMENTS


Fruits were collected and maintained by the Florida Department of Agriculture and Consumer Services, Division of Plant Industry (DPI), Bureau of Plant and Apiary Inspection. Many thanks to bureau chief Richard Clark, regional supervisors Terry Kipp and Debra Chalot, and the numerous bureau personnel who assisted in this project. Additional thanks are due to Calie Jenkins and Joyce Willis for their assistance to field studies, and to Sam Simpson for the use of his laboratory to hold fruit samples.

Additional collections were performed by the U.S. Department of Agriculture, Animal and Plant Health Inspection Service (USDA-APHIS). Thanks to John Thomas, Earl Wiley and Ralph Cooley.

Parasitoids and host flies were supplied by the DPI Caribbean Fruit Fly Mass Rearing Facility. Thanks to Don Harris, chief of the Bureau of Methods Development and Biological Control, and to Ed Bums, Suzanne Fraser, Mary Jo Hayes and other facility personnel for their assistance.

Tracy Austin, North Florida Research and Education Center, Quincy, Florida, supplied me with Caribbean fruit fly trapping data. Richard Brenner, David Milne, Jon Allen and Carlyle Brewster assisted in the preparation of contour maps. Yoav Gazit, Ali Harari and Kevi Vulinec cared for the laboratory cultures while I was away on field trips.


ii








Thanks to my supervisory committee members, Pat Greany, Howard Frank, Jonathan Crane and Richard Baranowski, for there useful suggestions. Special thanks to the committee chair, John Sivinski, for his assistance, support and patience.

Many thanks to Tim Holler, USDA-APHIS, for his assistance throughout the course of this dissertation. Among other things, Tim helped organize fruit collections, participated in the host plant surveys, and assisted in maintenance of laboratory cultures. Many of my insights regarded parasitoid distribution were born from hours of discussion with Tim during our travels together.

Thanks to my friends and colleagues at the Department of Entomology and Nematology and at the U.S. Department of Agriculture, Agricultural Research Service, who helped me maintain some degree of sanity.

Finally, thanks to my parents, whose many years of love, support and understanding have enabled me to achieve my goals.


iii














TABLE OF CONTENTS


PAe




A C K N O W LE D G M EN T S ...............................................................................................ii

ABSTRACT.................................................... ViI

CHAPTER 1 GENERAL INTRODUCTION: THE CARIBBEAN FRUIT FLY
AND ITS PARASITOIDS IN FLORIDA .................................................................. 1

CHAPTER 2 LITERATURE REVIEW: DISTRIBUTION, TEMPERATURE
TOLERANCE AND DIAPAUSE OF PESTIFEROUS TEPHRITID FRUIT
FLIES AND THEIR PARASITOIDS IN TROPICAL AND SUBTROPICAL
R E G IO N S ................................................................................................................. 5

D istribution and Population Dynam ics..................................................................... 5
H a w a ii. ................................................................................................................5
Anastrepha and Ceratitis in Tropical and Subtropical America ........................ 13
Bactrocera oleae in Southern Europe ................................................................. 18
Bactrocera tryoni in Australia ........................................................................ 18
C o n clu sio n ......................................................................................................... 19
Effects of Temperature and Occurrence of Diapause. ........................................... 20

CHAPTER 3 LARGE-SCALE DISTRIBUTION PATTERNS OF CARIBBEAN
FRUIT FLY PARASITOIDS IN FLORIDA ....................................................... 23

M aterials and M ethods......................................................................................... 24
F ruit Sam pling ............................................................................................. . . 24
Abiotic Environm ental D ata ........................................................................... 29
H ost Fly and H ost Plant D ata ......................................................................... 31
Statistical A nalysis ....................................................................................... . . 33
R e su lts ..................................................................................................................... 3 5
Distribution and Abundance of Parasitoids .................................................... 35
H ost P lant D ensity ....................................................................................... . 45


iv









Relationships With Environmental Factors...................................................... 52
D isc u ssio n ............................................................................................................... 5 5

CHAPTER 4 LOCAL TEMPORAL AND SPATIAL DISTRIBUTION
PATTERNS OF DIA CHASMIMORPHA LONGICA UDA TA AND
DORYCTOBRA CONAREOLA TUS IN AN AREA OF CO-OCCURRENCE ........... 72

M aterials and M ethods......................................................................................... 73
Fruit Sampling in 1996. .........................73
Analysis of Distribution Among Trees............................................................ 74
Temperature Measurements ........................................................................... 76
Comparisons of Parasitoid Abundance Among Years ..................................... 76
R e su lts .....................................................................................................................7 8
Comparison of Abundance Between LaBelle and Ft. DeNaud .............................78
D istribution A m ong T rees ..................................................................................79
Temperature Measurements ........................................................................... 81
C om parisons A m ong Y ears.............................................................................. 101
D isc u ssio n ............................................................................................................. 1 12

CHAPTER 5 EFFECTS OF TEMPERATURE ON IMMATURE
DEVELOPMENT, ADULT LONGEVITY AND OVIPOSITIONAL
ACTIVITY IN DIACHASMIMORPHA LONGICAUADATA ................................. 116

M aterials and M ethods........................................................................................... 116
Im m ature D evelopm ent .................................................................................... 116
Adult Longevity and Ovipositional Activity ...................................................... 118
R e su lts ................................................................................................................... 1 19
Im m ature D evelopm ent .................................................................................... 119
Adult Longevity and Ovipositional Activity ...................................................... 120
D iscu ssio n ............................................................................................................. 13 1

CHAPTER 6 LABORATORY REARING OF DORYCTOBRA CON
A R E O L A T U S ......................................................................................................... 134

M aterials and M ethods........................................................................................... 134
In se cts.............................................................................................................. 1 3 4
C ag e S etu p ...................................................................................................... 13 5
E xposure to H ost L arvae.................................................................................. 135
Immature Stages and Adult Emergence ............................................................ 136
L ife H isto ry T raits............................................................................................ 13 7
R esults and D iscussion ........................................................................................... 138


v








CHAPTER 7 BEHAVIORAL RESPONSE TO HOST CHEMICAL CUES BY
FEMALES OF DORYCTOBRACONAREOLA TUS ............................................... 143

M aterials and M ethods........................................................................................... 143
In se c ts .............................................................................................................. 14 3
E xperim ental D esign ........................................................................................ 144
Response Variables and Statistical Analysis ...................................................... 145
R e su lts ................................................................................................................... 14 6
D iscu ssio n ............................................................................................................. 14 7

CHAPTER 8 SUMMARY AND CONCLUSIONS..................................................... 152

APPENDIX NUMBERS OF SAMPLES COLLECTED AND INSECTS
EMERGING FOR VARIOUS SITES BY MONTH, YEAR AND FRUIT
T Y P E .................................................................................................................... 1 5 5

R E FE R E N C E S C IT E D ............................................................................................... 17 1

BIOGRAPHICAL SKETCH ....................................................................................... 183


vi













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

BIOGEOGRAPHY OF BRACONID PARASITOIDS OF THE CARIBBEAN FRUIT FLY, ANASTREPHA SUSPENSA (LOEW) (DIPTERA: TEPHRITIDAE), IN FLORIDA By

Avraham Eitam

May 1998

Chairman: Dr. John Sivinski
Major Department: Entomology and Nematology

Host fruits of the Caribbean fruit fly, including loquat, Surinam cherry, Cattley guava and common guava, were collected throughout central and southern Florida. Three species of braconid parasitoids were recovered. Diachasmimorpha longicaudata (Ashmead) was limited mostly to southern Florida, reaching higher latitudes along both coasts. Doryctobracon areolatus (Szepligeti) was common at most interior locations, but absent or rare along the coasts. Distribution of these two species overlapped only within a limited region, and only at LaBelle (Hendry Co.) did they commonly co-occur. Utetes anastrephae (Viereck) was widespread, but its abundance was inversely related with that of D. areolatus.

Absence of D. longicaudata was related with low temperatures, but was best explained by high variability of temperatures. Two hypotheses are proposed to explain the relationship between temperature and D. longicaudata distribution: (1) Low temperatures have a direct negative effect; (2) Variable or low temperatures adversely vii








affect host availability, which in turn has a negative effect on D. longicaudata. Evidence supporting each hypothesis is discussed.

Parasitism levels by D. longicaudata in loquat and common guava fruits were significantly related with the minimum and mean numbers, respectively, of Caribbean fruit flies captured in McPhail traps. Similarly, parasitism by U. anastrephae in loquat and Surinam cherry fruits was related with minimum fly numbers.

Parasitism levels of all species combined in loquat and Surinam cherry fruits was significantly related with densities of common guava trees. Parasitism by U. anastrephae in Cattley guava fruits was related with densities of Surinam cherry plants.

The apparent disappearance of D. areolatus from the southern Atlantic coast, where it was originally released, may be partially due to interspecific competition. Mechanisms proposed that may give D. longicaudata a competitive advantage include better ability to locate larvae within fruits, a longer ovipositor allowing greater access to hosts, higher fecundity, and an advantage in competition among larvae.

At LaBelle, parasitism by D. longicaudata in spring fruiting loquat and Surinam cherry was positively related with the preceding winter temperatures. A similar relationship was found for D. areolatus, but only in loquat fruits. A negative relationship between parasitism by D. areolatus and D. longicaudata was observed at the peak of the Surinam cherry fruiting season, suggesting that significant competition may occur.


viii













CHAPTER 1
GENERAL INTRODUCTION: THE CARIBBEAN FRUIT FLY AND ITS PARASITOIDS IN FLORIDA


The Caribbean fruit fly, Anastrepha suspensa (Loew), became established in Florida in 1965, quickly spreading throughout southern and central Florida (Weems 1966). In Indian River County, A. suspensa occurrence was linked primarily with the availability of various host fruits (Nguyen et al. 1992). Flies were collected mainly from loquat (Erioboaryajaponica (Thunb.)) during December-April, Surinam cherry (Eugenia uniflora L.) during May-June, and Cattley guava (Psidium cattleianum Sabine) during July-August, with greatest numbers reported from the latter two fruits. A population increase during one of the survey years in November-December was related to a second crop of Surinam cherry. In Dade County, weekly trap catches appeared to mirror temperature fluctuations (Hennessey 1994). Correlations of fly catches with rain and temperature together were significant for most years. Hennessey (1994) concludes that abiotic environmental factors and host availability interact to affect trapping frequency.

In an effort to control A. suspensa, several species of parasitoids were introduced to Florida (Baranowski et al. 1993). The first to be released was Doryctobracon areolatus (Parachasma cereus) (Szepligeti) (Hymenoptera: Braconidae: Opiine) (Baranowski and Swanson 1970). This is a widespread larval-pupal species, ranging from Mexico to Argentina (Wharton and Marsh 1978). An original stock of 7 males and 17 females from Trinidad was reared through 6-7 generations, and 45 males and 26 females were released


I





2


at Homestead in 1969 (Baranowski et al. 1993). Although it was recovered in large numbers the following summer, populations at this site have since declined (Baranowski et al. 1993). Small numbers persisted in the area at least until the occurrence of Hurricane Andrew in 1992 (Sivinski 1991, pers. comm.).

Studies indicate that D. areolatus abundance varies among locations in Florida. Sivinski et al. (1996) report it to be the dominant parasitoid in areas west of Lake Okeechobee. However, it was absent from their samples in southeastern Florida. Holler (unpublished data) failed to recover this species in a survey preceding augmentative releases of Diachasmimorpha longicaudata (Ashmead) along the central Atlantic coast of the state. However, subsequent intensive fruit sampling during 1993-1994 produced 16 D. areolatus from 7 trees (Denise Marshall, pers. comm.).

D. longicaudata, another larval-pupal opiine braconid, was introduced in 1972. This Indo-Philippine species was originally recovered from Bactrocera spp. (Clausen 1978). It has been utilized in the biological control of a wide range of tephritid hosts in various regions of the world (see Chapter 2). In contrast to the limited release of D. areolatus, D. longicaudata was released in large numbers in 21 counties throughout central and southern Florida (Baranowski et al. 1993). Based on reduced fly catches in subsequent years, it appeared to have had a significant impact upon host fly populations (Baranowski et al. 1993).

Two other exotic larval-pupal parasitoids, Aceratoneuromyia indica Silvestri (Hymenoptera: Eulophidae) and Trybiographa dad Weld (Hymenoptera: Eucoilidae), also were considered established (Baranowski et al. 1993). In addition two larval-pupal braconid parasitoids, Utetes anastrephae (Viereck) and Doryctobracon anastrephilum





3


(Marsh), were recovered in small numbers prior to the parasitoid releases. These were considered to have originally existed on Anastrepha interrupta Stone in the Florida Keys (Baranowski et al. 1993). Like D. areolatus, U. anastrephae is a wide ranging species, distributed south to Argentina (Wharton and Marsh 1978).

Sivinski et al. (1998) investigated the temporal dynamics of D. longicaudata and D. areolatus populations at LaBelle. D. longicaudata became more abundant, actually and relative to D. areolatus, as the season progressed, in all fruits except calamondin (Citrus mitis Blanco). Temperature best explained the fluctuations in relative abundance. However, with the exception of the autumn-winter decline of D. longicaudata in calamondin, results could also be explained by "counter-balanced competition" (cf. Zw6lfer 1971), where D. areolatus is superior to D. longicaudata in finding host patches, but inferior at exploiting hosts.

Augmentative releases of adult D. longicaudata apparently substantially suppressed A. suspensa populations at two locations in Florida (Sivinski et al. 1996). Similar results have been reported with releases of Diachasmimorpha tryoni (Cameron) for the suppression of the Mediterranean fruit fly, Ceratitis capitata (Wiedemann) in Hawaii and Guatemala (Wong et al. 1991, Sivinski et al. submitted). Inundative releases of D. longicaudata were being employed for A. suspensa control for several years in the central Atlantic coast region of Florida (Burns et al. 1996).

The objectives of this study were to determine the current distribution patterns and relative abundance of A. suspensa parasitoids in Florida, and identify factors affecting this distribution. These determinations could assist in the ongoing biological





4


control effort, by suggesting which parasitoid species should be employed in augmentative releases in various regions of the state.

The following chapter reviews literature on the distribution of tropical and subtropical tephritid fruit flies and their parasitoids in other regions of the world. Chapter 3 describes the geographic distribution of A. suspensa parasitoids in Florida, and the environmental factors associated with this distribution. Chapter 4 investigates temporal and spatial dynamics in an area of co-occurrence of parasitoid species within Florida. Subsequent chapters investigate biological attributes which may influence parasitoid distribution and abundance. Chapter 5 examines effects of temperature on D. longicaudata adults and immature stages in the laboratory. Chapter 6 describes life history traits of D. areolatus in the laboratory. Chapter 7 investigates the host location behavior of D. areolatus.

Generic names of opine braconids in this text are according to the recent revision by Wharton (1997).













CHAPTER 2
LITERATURE REVIEW: DISTRIBUTION, TEMPERATURE TOLERANCE AND DIAPAUSE OF PESTIFEROUS TEPHRITID FRUIT FLIES AND TIER
PARASITOIDS IN TROPICAL AND SUBTROPICAL REGIONS


Distribution and Population Dynamics


A great deal of literature has been published concerning distribution of tephritid fruit flies and their parasitoids in various regions of the world. The two regions most extensively studied in this regard have been Hawaii and tropical America. The objective of this review is to identify factors which may be important in explaining fruit fly and/or parasitoid distribution patterns.


Hawaii

Four species of adventive frugivorous tephritids occur in Hawaii. They are melon fly, Bactrocera cucurbitae (Coquillett), arrived in 1895; Mediterranean fruit fly, Ceratitis capitata (Wiedemann), in 1910; oriental fruit fly, B. dorsalis (Hendel), in 1945; and Malaysian fruit fly, B. latifrons (Hendel), in 1983 (Vargas et al. 1989).

Shortly after its arrival, C. capitata became a serious economic pest on various fruits throughout Hawaii (Back and Pemberton 1918). Subsequent to the arrival of the B. dorsalis, C. capitata became scarce at lower elevations but remained abundant in upland areas (Bess 1953). It was speculated that C. capitata had been competitively displaced by B. dorsalis, which was better adapted to the warmer climate of the lowlands (Haramoto


5





6


and Bess 1970). In guava fruits, B. dorsalis larvae suppressed the development of larvae of the C. capitata by an unspecified mechanism (Keiser et al. 1974). Since guava is a major host at low altitudes, this could contribute to the displacement of C. capitata in these areas. However, C. capitata has not been totally displaced at low elevations. Bess (1953) and Keiser et al. (1974) noted that it remained common in preferred hosts, such as coffee, regardless of elevation. During the winter months of 1966-1968 (when total infestation was low) it outnumbered B. dorsalis in guava (Haramoto and Bess 1970). Vargas et al. (1983a) studied C. capitata distribution on Kauai. Fly abundance was inversely related with elevation, and also with rainfall. Large numbers emerged from peach, loquat, sandalwood, and coffee. Highest populations were in areas containing scattered strands on feral coffee. In newly planted coffee fields in lowland Kauai, C. capitata dominated B. dorsalis by the end of each of three seasons (Vargas et al. 1995). The authors suggest that infestation of an earlier stage of fruit ripeness and faster development of C. capitata larvae in coffee reduce competitive interactions with B. dorsalis larvae. Furthermore, they suggest that absence of large overstory trees and a scarcity of alternate hosts may limit B. dorsalis abundance in monoculture coffee fields. In studies on Oahu, C. capitata was found to occur in larger numbers than B. dorsalis in feral coffee, but was less common on other hosts (Harris and Lee 1986, 1987). The abundance of coffee berries and distribution of other fruits apparently influenced fluctuations in C. capitata populations. Rainfall had an indirect effect on C. capitata dynamics by inducing coffee fruiting (Harris and Lee 1986). In the urban areas of Oahu, there was apparently a more direct effect of rainfall. Medflies were more common in dry,





7


leeward areas, although the same host fruits were grown there as in wet, windward areas (Harris and Lee 1987).

Population cycles of B. dorsalis have been associated with fruiting of common and strawberry guava (Newell and Haramoto 1968, Vargas et al. 1983b). Although abundance is negatively related with elevation on Kauai, it appears that the factor limiting distribution at high elevations is the relative scarcity of hosts (Vargas et al. 1983b). Similarly, high numbers of B. dorsalis in wet windward areas vs. dry leeward areas, and outside vs. inside production areas, corresponds with concentrations of wild guava (Vargas et al. 1989, 1990).

The major hosts of B. cucurbitae in Hawaii include tomato and various species of wild and cultivated cucurbits (Harris et al. 1986). Fly abundance on Kauai is negatively related with elevation and rainfall. These relationships may be explained by host plant distribution: hosts are not found above 300 m, and grow better in drier areas (Harris et al. 1986). On Molokai, fly distribution is strongly related to that of the feral host bittermelon (Harris and Lee 1989). Similarly, abundance of B. cucurbitae in dry leeward areas vs. wet windward areas on Kauai is related to the distribution of bittermelon and spiny cucumber (Vargas et al. 1989). More B. cucurbitae were captured inside production areas, again related to abundance of host plants (Vargas et al. 1989, 1990).

B. latifrons develops on a variety of solanaceous and cucurbitaceous plants (Liquido et al. 1994). Although B. cucurbitae is the primary fruit fly on most species in these families, B. latifrons appears to outcompete other fruit fly species on several host plants that inhabit disturbed, abandoned fields and less managed ranch lands (Liquido et





8


al. 1994). Populations are apparently affected by both temperature and rainfall. Liquido et al. (1994) suggest that high rainfall excludes this fly from the windward side of Hawaii.

Following the arrival of C. capitata, a project was undertaken to introduce fruit fly parasitoids to Hawaii. F. Silvestri (during 1912-1913) and D.T. Fullaway and J.C. Bridwell (in 1914) released several species, 5 of which became established: the opiine braconids Diachasmimorpha tryoni (Cameron), Opius humilus Silvestri, and Diachasmimorpha (Biosteres) fullawayi (Silvestri), the eulophid Tetrastichus guffardianus Silvestri, and the chacidid pupal parasitoid Dirhinus giffardii Silvestri (Clausen et al. 1965, Gilstrap and Hart, 1987). All species were from Africa, with the exception of D. tryoni from Australia.

Initially, 0. humilus became the dominant species, reaching maximum levels of parasitism in 1915. D. tryoni was dominant from 1916 onward, with maximum parasitism recorded in 1918. Total parasitism from 1914 to 1933 ranged between 24.9-56.4% (Willard and Mason 1937). 0. humilus disappeared from Oahu in the late 1930s, but Clausen et al. (1965) reported it as abundant in the Kona section of the island of Hawaii, equaling or exceeding D. tryoni in coffee. Interestingly, 0. humilus is not reported in more recent literature. The displacement of 0. humilus by D. tryoni may be due to a competitive advantage by the strongly mandibulate larvae of the latter species (Pemberton and Willard 1918).

After its establishment, T. gifardianus parasitized up to 25.3% of C. capitata larvae, averaging 6.3% between 1914-1933 (Willard and Mason 1937). Subsequent to Clausen et al. (1965), it had not been reported until Ramadan and Wong (1990) found it to be abundant in the Kula area of Maui. Purcell et al. (1994) found that while this





9


species is absent from guavas collected on the tree, it is common in fruit remaining on the ground for 4-9 days. Purcell (submitted) reports that another eulophid, Aceraneuromyia indica Silvestri, is established on all major islands, but appears to be less abundant than T. giffardianus.

Until the early 1950s, D. fullawayi was readily recovered from coffee and peach (Bess et al. 1961). Haramoto and Bess (1970) found this species only in coffee plantations in Kona, Hawaii. It has not been reported since.

Between 1947-1952, many species of parasitoids were introduced to Hawaii for the control of B. dorsalis (Clausen et al. 1965). Of these only four species of opiine braconids became permanently established: Fopius (Biosteres) arisanus (Sonan), Diachasmimorpha longicaudata (Ashmead), Fopius (Biosteres) vandenboschi (Fullaway), and Psyttalia incisi (Silvestri).

There was an interesting succession of parasitoids between 1948-1950. Initially, D. longicaudata was the dominant species. During the late summer and fall of 1949, F. vandenboschi increased in abundance, and by the end of the year had become far more numerous than D. longicaudata (Bess et al. 1950). During 1950, F. arisanus increased dramatically, and on Oahu constituted 99.4% of the total parasitism in December (van den Bosch et al. 1951). The total parasitism also increased, to approximately 80% (van den Bosch et al. 1951). F. arisanus has remained the dominant parasitoid ever since.

Several factors may explain this successive displacement. F arisanus attacks the egg of its host (van den Bosch and Haramoto 1951), F vandenboschi first-instar larvae, and D. longicaudata second and third-instar larvae. While all eggs are accessible to F arisanus, some larvae may escape parasitism by burrowing into the fruit pulp, especially





10


in large fruit (Sivinski 1991). This phenomenon would increase with older larvae, thus putting D. longicaudata at a disadvantage. Additionally, F. arisanus larvae inhibit the development of F. vandenboschi and D. longicaudata, and those of F. vandenboschi inhibit the development of D. longicaudata (van den Bosch and Haramoto 1953). This inhibition is apparently by means of physiological suppression and not physical injury (van den Bosch and Haramoto 1953). The early predominance of D. longicaudata was apparently enhanced by the disproportionate release of large numbers of individuals of this species (van den Bosch et al. 1951).

Note that Palacio et al. (1991) found no evidence of physiological suppression in a study on competition among F. arisanus, D. longicaudata and Fopius (Biosteres) persulcatus Silvestri. Instead, they found that D. longicaudata was a superior competitor to both F. arisanus and F. persulcatus, with F. persulcatus being superior to F. arisanus, indicating physical competition among first-instar larvae. They further reported that D. longicaudata did not discriminate between parasitized and unparasitized hosts, while F. persulcatus avoided superparasitism. However, Lawrence et al. (1978) demonstrated that D. longicaudata does avoid superparasitism when provided with large numbers of hosts.

F. arisanus, D. longicaudata, and F. vandenboschi develop not only on B. dorsalis, but also on C. capitata. The reduced abundance of C. capitata in the early 1950s was partially attributed to the effects of these parasitoids, particularly F. arisanus (Bess 1953). P. incisi does not develop on C. capitata (Stark et al. 1994).

D. tryoni does not develop on B. dorsalis, except in cases of multiparasitism involving D. longicaudata (Ramadan et al. 1994a). However, it develops on two species





11


of gall-forming tephritids, the eupatorium gall fly, Procecidochares utilis Stone, and the lantana gall fly, Eutreta xanthochaeta Aldrich (Haramoto and Bess 1970).

F. arisanus is the dominant parasitoid of both B. dorsalis and C. capitata in Hawaii. However, other species may constitute a large part of the total parasitism under certain circumstances. Wong et al. (1984) and Wong and Ramadan (1987) studied the parasitoid fauna on both species of fruit flies in the Kula area of Maui. In these studies, fly pupae were not separated, so the various parasitoid species could not be attributed to a specific fly species. D. longicaudata and D. tryoni were quite common on loquats and peaches, occasionally surpassing F. arisanus in abundance. For example, D. longicaudata accounted for 34.9% of the total parasitism in 1979 from peaches, and 33.4% in 1984 from loquats, and D. tryoni accounted for 32.7% in 1980 from loquats. P. incisi and F. vandenboschi accounted for 1.6% and 0.2 of the total parasitism, respectively.

At a site at 1200 m elevation on Hawaii island, D. tryoni was the dominant species, and often the only one recovered (M. Purcell pers. comm.). Dominance at high altitudes may be related to an observation by Pemberton and Willard (1918) that mature D. tryoni larvae enter a winter diapause within host puparia.

Several additional studies have reported relative abundance of B. dorsalis parasitoids. Vargas et al. (1990) reported that in ripe fruit in an agricultural area, D. longicaudata and P. incisi constituted 3.9-5.5% and 0.4-2.3% of the total parasitism, respectively. Stark et al. (1991) determined the abundance of parasitoids in commercial guava by canopy fogging. D. longicaudata, P. incisi, and F. vandenboschi accounted for 9-10%, 2-10%, and 0.25-1% of the total parasitism, respectively. Vargas et al. (1993)





12


report that these three species were more common in orchards than in wild guava. They suggest that two possible factors contributing to this observation may be high tree densities and abundance of rotting fruit in commercial guava orchards. Vargas et al. (1993) point out that fruit type influenced parasitoid abundance. F vandenboschi represented 8.8% of the total parasitoids collected from passion fruit in 1988. Small fruits such as Surinam cherry and false kamani produced many P. incisi, while D. longicaudata was often common in mango (32.2% in 1988).

Purcell et al. (1994) sampled guava fruit from the tree and the ground. They found that D. longicaudata abundance increased as fruit aged on the ground, and P. incisi was recovered only from fruit on the ground at least 4 days. This suggests that these parasitoids forage on fruit on the ground, and sampling fruit only from the tree would underestimate their abundance. F vandenboschi abundance (less than 3% of total parasitoids) was unaffected by fruit ripeness.

Vargas et al. (1995) studied abundance of Mediterranean and oriental fruit flies and their parasitoids in newly-planted coffee fields. Although parasitism of C. capitata by F. arisanus was apparently density-dependent, low parasitism (33.1-37.6%) was observed. Interestingly, F arisanus parasitized a greater percentage of C. capitata, the more common host, than B. dorsalis. On Oahu, F arisanus appeared to be inefficient in parasitizing hosts at low population densities (Harris and Lee 1987).

The opiine braconid Psyttalia fletcheri (Silvestri) was introduced in 1915-1916 for control of B. cucurbitae (Clausen et al. 1965). It is the only significant parasitoid of B. cucurbitae in Hawaii (Nishida 1955). This species was more common on the wild hosts Momordica balsamina L. and M. charantia L. than on cultivated hosts, with parasitism





13


levels of up to 50% during favorable seasons (Nishida 1955). Parasitism was highest in winter and lowest in summer (Nishida 1955). Larvae in vines were more highly parasitized than larvae in fruits, possibly because in the latter they could escape parasitism by penetrating deeply into the pulp (Nishida 1955). This may also explain the low parasitism on large cultivated fruits. Parasitism in cultivated fruits was higher in very weedy fields, suggesting that P. fletcheri favors weedy situations (Nishida 1955). Harris and Lee (1989) suggest that the absence of P. fletcheri from Molokai may be due to unfavorable high winds on that island.

Liquido et al. (1994) report very low (less than 1%) parasitism of B. latifrons by D. longicaudata and Tetrastichus sp.


Anastrepha and Ceratitis in Tropical and Subtropical America


The genus Anastrepha includes 184 described species ranging from the southern United States to northern Argentina (Aluja 1994). At least 54 species occur in Panama (Stone 1942) and 23 in Mexico (Aluja et al. 1987). Biological knowledge is basically restricted to seven economically important species: fraterculus, grandis, ludens, obliqua, serpentina, striata, and suspensa (Aluja 1994).

Several studies of Anastrepha abundance have been conducted in Chiapas, southern Mexico. Celedonio-Hurtado et al. (1995) trapped flies in orchards of various fruit species. Fruit fly species composition varied among orchards, with 1 or 2 predominant species representing 43-86% of all individuals. For example, in sapodilla, Achras zapota L., 86% of all flies trapped were A. serpentina (Wiedemann), while in chalum, Inga micheliana Harms, 66% were A. distincta Greene and 25% A. ludens (Loew). Rainfall could not explain population fluctuations, and the authors conclude that





14


host fruit availability is the most important factor affecting adult populations. A. obliqua (Macquart) and A. ludens are the predominant species in mango, with A. obliqua being more common at lower elevations and A. ludens at higher elevations (Aluja et al. 1987, 1990, 1996). In another study conducted in a coffee producing area of the same state, A. ludens was the most abundant species with 60% of trapped flies, followed by A. distincta and A. fraterculus (Wiedemann) with 22 and 12%, respectively (Malo et al. 1987). The occurrence of the latter two species was related to the abundance of Inga spp., the main host of A. distincta, and coffee, a minor host of A.fraterculus, in the study area.

Studies in other countries showed similar tendencies. In Costa Rica, A. obliqua was associated with mango and other Anacardiaceae, A. striata Schiner with guava and other Myrtaceae, and A. serpentina with species of Sapotaceae (Jir6n and Hedstrom 1988, 1991). Soto-Manitiu and Jir6n (1989) found that the maximum abundance of each species coincides with the fruiting season of their respective host plants. Most A. obliqua emerge just after first rains, coinciding with the mango fruiting season. In citrus orchards in Belize, the seasonal increase in numbers of A. ludens trapped was derived mainly from infestations in grapefruit (Houston 1981).

In Brazil population dynamics of A. fraterculus has been related to host fruit availability (Malavasi and Morgante 1981). Nascimento et al. (1982) report that A. obliqua was predominant in citrus orchards, and A. fraterculus in localities with tropical hosts, especially guava. While the occurrence of flies in citrus was related to host availability, no such relationship was observed with tropical hosts. Trapping was related to mean and minimum temperature and relative humidity. Fehn (1982) studied the population dynamics of Anastrepha spp. in peach orchards at three locations in two





15


seasons. He found relationships with various meteorological factors, including temperature, relative humidity, rainfall and wind velocity, at some locations and seasons but not others. However, he suggests that availability of alternative hosts may be the principal factor affecting population dynamics.

C. capitata invaded Costa Rica in 1955 and has since spread to all of Central America (references in Wharton et al. 1981), and into South America to Brazil (Aguiar and Menezes 1996). It comprised 5 and 19% of fruit flies collected in Costa Rica and Brazil, respectively (Jiron and Hedstrom 1988, Aguiar and Menezes 1996).

Several species of parasitoids were released for the control of C. capitata in Central America (Gilstrap and Hart 1987). Of these, three species--D. longicaudata, F. arisanus and the eulophid A. indica-were recovered by Wharton et al. (1981) in Costa Rica. In this study, D. longicaudata and F. arisanus were the dominant parasitoids of C. capitata, while A. indica and D. longicaudata were the most common on Anastrepha spp. Native parasitoids occurred in much smaller numbers. For example, parasitism of Anastrepha spp. by the opiines Doryctobracon areolatus (Szepligeti) and Utetes anastrephae (Vier.) was only 0.2 and 0.05%, respectively. In a later study, Jir6n and Mexzon (1989) report that D. areolatus was the most abundant and widespread species. In Guatemala, D. longicaudata was reported to be the most common parasitoid of C. capitata, while that of Anastrepha spp. was Doryctobracon crawfordi (Viereck), followed by D. areolatus and U anastrephae (Eskafi 1990). The combined parasitism in this study was very low (<2%) for all fruits except Surinam cherry with 8% parasitism.

Various exotic fruit fly parasitoids were introduced into Mexico in the 1950s (Jimenez-Jim6nez 1956, 1958). Of these, D. longicaudata and A. indica were established





16


(Clausen 1978). Several systematic surveys of Anastrepha parasitoids were subsequently conducted, producing very different results concerning the relative abundance of parasitoid species. In an area of mixed cultivation in the State of Chiapas in southern Mexico, Aluja et al. (1990) found that the most abundant parasitoid was D. longicaudata. However, in a native tropical community in the State of Veracruz, D. areolatus and U. anastrephae represented 59 and 17%, respectively, of the total parasitism, while D. longicaudata was not recovered at all (Hernandez-Ortiz et al. 1994). Lopez et al. (submitted) confirmed that D. areolatus is the most common species in Veracruz, representing 43% of all parasitoids recovered from various habitats. It also had the widest host breadth of all parasitoid species. In an earlier study in the State of Nuevo Le6n in northeastern Mexico, Gonzalez-Hernandez and Tejada (1979) reported that the most common parasitoid was D. crawfordi, followed by D. areolatus. Interestingly, in Veracruz D. crawfordi was common only in citrus (Lopez et al., submitted).

Exotic parasitoids also were introduced into other locations in the Americas for the control of Anastrepha species. D. longicaudata was introduced to Trinidad in 1974 (Bennett et al. 1977). The following year it was the most common parasitoid recovered, surpassing the native D. areolatus. In Argentina, D. longicaudata and A. indica were reported as established (Ovruski and Fidalgo 1994).

Several studies in Brazil report that D. areolatus was by far the most abundant parasitoid of Anastrepha species. Leonel et al. (1995) found that 70% of parasitoids emerging from samples collected in 10 states were D. areolatus. The alysiine Asobara anastrephae (Muesebeck) was the second most common species with 19% of the total parasitism, and U. anastrephae the third most common with 10%. In the state of Sdo





17


Paulo alone, D. areolatus and U. anastrephae constituted 84 and 6% of all parasitoids, respectively (Leonel et al. 1995). In Itaguai, Rio de Janeiro, 89% of all parasitoids collected were D. areolatus, with U anastrephae accounting for an additional 8% (Aguiar-Menezes and Menezes 1997). In Amazonas State, D. areolatus was found to be the dominant parasitoid in rural locations while Opius sp. nr. bellus dominated in urban areas (Canal D. et al. 1994, 1995). Finally, D. areolatus was the dominant parasitoid of Anastrepha zenildae Zucchi in the State of Rio Grande do Norte (Araujo et al. 1996).

D. areolatus also was the most common parasitoid reported from Venezuela, accounting for 33% of the parasitism (Katiyar et al. 1995). U anastrephae was the fourth most common species with 7%. Ovruski (1995) reports low levels of parasitism from Tucumin province, Argentina, with D. areolatus emerging from less than 2% of Anastrepha spp. puparia. Earlier studies by Nasca (1973) and Fernandez de Arioz and Nasca (1984) also reported D. areolatus (as Opius tucumanus or Doryctobracon tucumanus) from the same province.

Sivinski et al. (1997) analyzed the distribution of parasitoids of Anastrepha spp. within tree canopies in Mexico. Several tendencies were reported. Parasitism by U. anastrephae was observed only in a narrow range of small host fruits. The efficiency (proportion of larvae attacked in a fruit) of D. longicaudata compared to that of other parasitoids increased with fruit size. Parasitism by D. areolatus decreased during fruiting periods of individual trees as the season changed from rainy to dry. Negative relationships in parasitism were observed between D. areolatus and U. anastrephae, while the introduced D. longicaudata and native D. crawfordi tended to overlap.





18


Bactrocera oleae in Southern Europe


The olive fly, Bactrocera oleae (Gmelin) is native to Africa and currently distributed throughout the Mediterranean basin and the Middle East (Clausen 1978). The braconid Psytalia concolor (Silvestri) was introduced to Italy from North Africa in 1914, and again in 1917-1918, 1923 and 1934 (Clausen 1978). It was also established in France and Greece (Clausen 1978). Inundative releases have been performed at various locations and have proven to be quite successful in reducing fly populations (e.g., Monastero and Delanoue 1966, Kapatos et al. 1977). While it is established in southern Italy, attempts at establishment in more northern regions have failed (e.g., Raspi and Loni 1994).



Bactrocera tryoni in Australia


The Queensland fruit fly, Bactrocera tryoni (Froggatt), is native to Australia. Toward the southern fringe of its permanent distribution, there was a significant correlation between summer rainfall and peak fly numbers (Bateman 1972). The effect was thought to be mediated through a reduction in fecundity and immigration, and high mortality among adults emerging through dry soil in dry years.

Several species of exotic parasitoids were introduced to Australia (Snowball et al. 1962). Only F. arisanus persisted on the Australian mainland and D. longicaudata on Lord Howe Island, even though both species were initially established at both locations (Snowball and Lukins 1964). At most locations, native parasitoids were uncommon relative to F. arisanus. Snowball and Lukins (1964) suggest that low winter temperatures may limit the distribution of F. arisanus in southern Australia. Snowball (1966)





19


concludes that factors other that host availability may account for lower and more variable parasitism at higher latitudes, again suggesting some temperature effect.



Conclusion


Factors affecting distribution of fruit flies and their parasitoids could be put into three main categories: abiotic factors, host availability and competition.

The most commonly mentioned abiotic factors are temperature and precipitation. Temperature may be important at limiting distribution at high latitudes, as was suggested for F. arisanus in southern Australia. Winter temperatures may also limit parasitoid distribution in Florida, given that these parasitoids originate in tropical regions and may not be adapted to cold temperatures.

The factor most commonly mentioned as affecting distribution and abundance of tropical fruit flies and their parasitoids is host fruit availability. In addition to many studies from Hawaii and tropical America noted above, Tan and Serit (1994) reached similar conclusions regarding B. dorsalis in Malaysia. Even in cases in which population dynamics appear to follow temperature changes, several authors have suggested an indirect effect of temperature, through its influence in host availability.

Interspecific interactions among parasitoids may be very complex. Examples of successive replacement of parasitoid species in Hawaii have been regarded as classic examples on competitive displacement. Mechanisms suggested have involved competition among larval stages within fruit, including either physical interaction among larvae or physiological suppression. Additionally, parasitoids attacking early stages replaced parasitoids attacking later stages. It was suggested that earlier stages are more





20


exposed to parasitoids because they are situated closer to the fruit surface, thus giving the former species an advantage. This mechanism may be irrelevant in Florida, as the three parasitoid species present in the state, namely D. areolatus, D. longicaudata and U. anastrephae, apparently all attack late-instar larvae.


Effects of Temperature and Occurrence of Diapause


The distribution of tropical fruit flies in warm temperate climates can be limited by temperature. Therefore, the determination of low temperature tolerance is important. Messenger and Flitters (1954) used environmental chambers to simulate the climates of locations in the continental United States. They determined that C. capitata, B. cucurbitae, and B. dorsalis could successfully reproduce in most of Florida and along the Gulf coast, and possibly southern California. Flitters and Messenger (1965) stated similar conclusions for A. ludens. Meats (1981) and O'Loughlin et al. (1984) determined that the Queensland fruit fly, Bactrocera tryoni (Froggatt), could establish permanent, lowdensity populations in southern Australia.

Levya-Vazquez (1988) used a degree-day model to estimate lower thresholds and thermal constants for A. ludens. The lower thresholds for the various immature stages ranged from 9.4-14.1*C. Thomas (1997) found that pupal duration in the field closely fit this laboratory-based model. The puparial stage may be prolonged up to three months in the winter, but there was no evidence of a winter diapause. Larval development time was variable and did not agree well with the model.

Christenson and Foote (1960) reviewed the occurrence of diapause in fruit flies. While diapause is typical of most North American Rhagoletis spp., most tropical and subtropical species are not known to undergo diapause.





21


Prescott and Baranowski (1971) determined the temperature tolerances for A. suspensa. Eggs failed to hatch below 12'C or above 33*C. No emergence was observed at 10 and 12*C, although pupae were still viable at 12*C when the experiment was terminated. The calculated development threshold was 10*C and the optimal development temperature approximately 25*C for all immature stages.

Ashley et al. (1976) studied adult emergence of A. suspensa and D. longicaudata between 22-32*C. Both flies and parasitoids had high levels of mortality above 28*C. In contrast Darby and Kapp (1934) report that A. ludens has greater tolerance than its parasitoid D. crawfordi at both high and low temperatures. No emergence was observed for D. crawfordi at 12 and 30*C and for A. ludens at 10 and 31*C.

Loni (1997) studied the effects of temperature on the development of P. concolor. Adult emergence was greatest between 18-25*C, markedly reduced at 15 and 280C, and zero at 13 and 33*C.

Pemberton and Willard (1918) recorded diapause for D. tryoni and D. fullawayi. Darby and Kapp (1934) observed delayed emergence in several individuals of D. crawfordi. Clausen et al. (1965) reported diapause in D. longicaudata strains collected from areas having cool winters. Ashley et al. (1976) observed an increase in delayed emergence of D. longicaudata larvae at the lowest temperature (220C) and likewise with low moisture concentration. Finally, Aluja et al. (submitted) recorded diapause in Mexican populations of D. areolatus, D. longicaudata and U. anastrephae, and also in Aganaspis pellenaroi (Brethes) and Odontosema anastrephae Borgmeier (Hymenoptera: Eulophidae).





22

In conclusion, temperature tolerances could determine the limits of distribution for fruit flies and their parasitoids. Laboratory studies of temperature effects on Caribbean fruit fly parasitoids could help ascertain the influence of temperature on their distribution in Florida.













CHAPTER 3
LARGE-SCALE DISTRIBUTION PATTERNS OF CARIBBEAN FRUIT FLY PARASITOIDS IN FLORIDA


Studies from many geographical regions have indicated that distribution of fruit flies and their parasitoids may be affected by a variety of factors, including temperature, precipitation, host fruit availability, and interspecific competition (Chapter 2). Parasitoids may themselves be affected by host fly abundance.

In addition to mean or extreme temperature and precipitation, parasitoids may be influenced by the variance of these factors. In particular, the variability of abiotic factors among months could affect the temporal dynamics of hosts. Fruit yield is highly dependent on environmental factors, and can be adversely affected by temperatures or rainfall that are periodically higher or lower than optimal (Petr 1991, Raper and Kramer 1983). High variance in these abiotic factors would lead to greater variability in the temporal availability of hosts. Parasitoid species may differ in their ability to survive through periods of low host abundance.

Three species of parasitoids are commonly recovered from the Caribbean fruit fly, Anastrepha suspensa (Loew), in Florida: Diachasmimorpha longicaudata (Ashmead), Doryctobracon areolatus (Szepligeti) and Utetes anastrephae (Viereck) (Hymenoptera: Braconidae: Opiine). Studies conducted at several locations in Florida suggest that distribution of parasitoid species may differ at various sites (see Chapter 1).


23





24


The objective of this chapter was to determine the parasitoid distribution throughout central and southern Florida, and identify abiotic and biotic factors possibly influencing this distribution.




Materials and Methods


Fruit Sampling


Host fruits of A. suspensa were collected in 23 towns in central and southern Florida (Figure 3-1). Sample sites were chosen to represent various regions of the peninsula. Thus samples were collected from 5 sites along the Atlantic coast from Melbourne (28.10 N) to Miami (25.80 N), 7 sites along the Gulf of Mexico coast from Tampa (28.0* N) to Naples (26.10 N), and 11 sites in the interior from Dade City (28.4* N) to Belle Glade (26.70 N) and LaBelle (26.80 N). Interior sites were situated along various north-south routes, e.g., US 17 (Lakeland, Wauchula and Arcadia) and US 27 (Haines City, Lake Wales, Lake Placid and Belle Glade). Clewiston, on the south-west coast of Lake Okeechobee, and Immokalee, 37 km south of LaBelle, were not sampled because of previous mass releases of D. longicaudata at these locations (Sivinski et al. 1996). Note that there are no interior towns south of those indicated due to the presence of the Everglades.

Sampling was not always limited to the town indicated, and often included collections in adjacent towns. Many Melbourne samples were actually collected in southern Brevard County, and St. Petersburg samples were collected throughout Pinellas County. Lake Placid samples include some from Sebring, and Punta Gorda includes





25


Atlantic

Dade City. Ocean H St. Cloud Melbourne
Tampa L e0 n

Lake Wales St. Petersbu

Ft. Pierre
Bradenton * Wauchula Okeec bee
Lake PI
GlLak
Venice Okeechobee
Punta Go a W. P m Beach

0 Belle Glade
LaBelle
Ft. yes

Ft. Lauderdalc
0 Naples Gulf of

Mexico Miami


Figure 3-1. Fruit sample collection sites.





26


samples from Port Charlotte. On the other hand, two large counties had more than one collection site. Haines City, Lakeland and Lake Wales are all in Polk County, while both West Palm Beach and Belle Glade are in Palm Beach County.

Samples were collected in August 1994, and monthly from January-September 1995. Additional monthly samples were collected at Melbourne, Bradenton, Venice and Okeechobee from March-May 1996 and at St. Petersburg in May 1996. Sampling was not performed from October through December, because at the majority of sites primary host fruits are uncommon during this period in most years (Tim Holler, pers. comm.). All samples were collected within a single week each month. Sampling at Ft. Pierce was conducted by Tim Holler, USDA-APHIS-PPQ, in March and May 1993, prior to augmentative releases of D. longicaudata. Sampling at most sites during 1994 and 1995 was performed by the Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Bureau of Plant Inspection. Sampling at other sites during 1994 was by USDA-APHIS.

Fruits sampled included loquat (Eriobotrya japonica (Thunb.)), Surinam cherry (Eugenia uniflora L.), Cattley guava (Psidium cattleianum Sabine) and common guava (Psidium guajava L.). Loquat samples were collected from January-April, most Surinam cherry samples from April-June, most Cattley guava from July-August, and most common guava from August-September. Note that additional fruiting periods may occur, especially at southern sites. Up to 12 samples were collected for each site during a single month. Each sample included fruits collected from a single tree. Fruits collected were ripe and usually without holes caused by beetles or exiting larvae. They were preferably collected from the tree, but occasionally supplemented with fruits from the ground.





27


Numbers of samples and total fruits collected at various sites are detailed in Table 3-1. Total sample numbers varied widely among sites, ranging from 17 samples collected at Ft. Pierce to 102 at LaBelle. The number of samples collected was dependent primarily on availability of host fruit. Note that the efficiency of fruit sampling may have varied among sites, as it was often conducted by different personnel at various sites. Cattley guava was the least commonly collected host, with 83 samples, compared with 549, 377 and 290 samples of loquat, Surinam cherry and common guava, respectively.

Fruits were placed within a bucket upon a metal screen. The bucket had holes for ventilation. It was covered with a plastic lid and its inside was lined with cloth to prevent entry of insects after fruit collection and escape of insects emerging from the fruit sample. Moist fine vermiculite (ca. 15 ml water per 100 cm3 vermiculite) was placed at the bottom of the bucket. Mature fruit fly larvae exited the fruit and pupated in the vermiculite.

At the end of each sampling week, buckets were collected from various locations and transported to Gainesville. Buckets were maintained at 25.50 C, except in 1995 when they were kept in a warehouse at ambient temperatures. Puparia were sifted from the vermiculite 13-15 days after fruit collection, and transferred to 250 ml plastic containers. These containers were initially covered with a solid lid, which was replaced after ca. one week with a screened lid. This was done to assure that the vermiculite did not dry out, but was also not so moist as to allow development of fungi. Containers were maintained at 25.50 C and ambient humidity.





28


Table 3-1. Total number of host fruit samples and total fruits collected at various sites.

Site Loquat Surinam cherry Cattley guava Common guava
Samples Fruits Samples Fruits Samples Fruits Samples Fruits Arcadia 35 1162 11 1076 3 94 10 51
Belle Glade 8 139 31 1202 1 22 31 100
Bradenton 32 873 24 1433 1 1 7 59
Dade City 29 1017 0 0 0 0 6 13
Ft. Lauderdale 9 321 30 1469 2 50 13 47
Ft. Myers 27 854 14 1090 19 510 5 50
Ft. Pierce 6 610 10 1550 0 0 1 22
Haines City 22 567 9 349 0 0 0 0
LaBelle 39 1272 28 2021 5 210 30 226
Lakeland 35 663 16 614 4 100 26 122
Lake Placid 37 1261 19 1534 3 98 11 75
Lake Wales 21 713 18 916 0 0 29 192
Melbourne 42 1241 21 1181 4 140 17 166
Miami 3 72 12 553 2 27 31 171
Naples 26 630 23 1129 24 806 3 --Okeechobee 21 713 20 1385 0 0 5 22
Punta Gorda 25 645 19 1160 10 353 0 0
St. Cloud 12 351 16 621 0 0 21 191
St. Petersburg 29 685 11 736 0 0 10 28
Tampa 29 561 0 0 0 0 8 --Venice 15 441 20 1615 0 0 0 0
Wauchula 40 1494 15 1185 2 61 22 133


W. Palm Beach 7 148 10 345 3 41 5 22


W. Palm Beach 7


148 10 345 3


41 5 22





29


Flies and parasitoids emerged within the containers, and were counted when no more live insects were observed. Parasitism levels for each species were calculated for each sample as the ratio between the number of parasitoids of the relevant species emerging and the sum of all flies and parasitoids emerging. This assumes that neither the flies nor the parasitoids diapause, and that mortality levels of the fly pupae and immature parasitoids are similar. In Florida, emergence rates of pupae held indoors under controlled temperature and humidity are typically ca. 90% (Sivinski, pers. comm.), so that it is unlikely that a significant proportion of parasitoids undergo diapause. In contrast, diapause appears to be quite prevalent among parasitoids held under seminatural conditions in Mexico (Aluja et al., submitted). Note that true levels of parasitism are underestimated, because fruit removed from the field include host eggs and larvae that may have been parasitized if left in place. However, comparisons of parasitism levels as measured should reflect relative abundance. Abiotic Environmental Data


Temperature and precipitation data were obtained from the Southeast Regional Climate Center, Columbia, South Carolina. Weather stations exist in most of the towns included in the study. However, there are no data for the vicinity of Dade City and Haines City. Data from Winter Haven and Avon Park were used to represent Lake Wales and Lake Placid-Sebring, respectively. The following variables were obtained: mean annual temperature, mean minimum temperature for the coldest month of the year, mean maximum temperature for the warmest month of the year, extreme annual minimum and maximum temperatures, and annual precipitation (Table 3-2).





30


Table 3-2. Mean precipitation and temperature values for towns in central and southern Florida, for the years 1980-1996. Data were obtained from the Southeast Regional Climate Center, Columbia, South Carolina.


Town Annual Mean
precipitation annual
(mm) temp (*C)


Arcadia Avon Park Belle Glade Bradenton Ft. Lauderdale Ft. Myers Ft. Pierce LaBelle Lakeland Melbourne Miami Naples Okeechobee Punta Gorda St. Cloud St. Petersburg Tampa Venice Wauchula W. Palm Beach Winter Haven


1341 1263 1290
1412
1672 1390
1403 1310 1320
1243
1509 1325
1224
1309 1268 1270
1133
1140 1300 1609 1286


22.2 22.3 22.8 22.8
24.4 23.9 23.0 23.3 23.1 22.5
24.8 23.8 23.1 23.5
22.6 23.3 22.7 22.9 22.8
24.1 23.0


aFor the coldest month of the year. "For the warmest month of the year. Temp = temperature.


Mean minimum temp (*C)a
8.0 7.6 9.6 9.0 13.9 11.3
9.3 9.0 9.1 9.2
14.3 11.3
8.3
10.2
8.6 11.3 9.2
8.8 9.7
12.8
9.2


Extreme minimum temp (*C)
-3.0
-3.0
-0.8
-1.8
2.9 0.7
-1.8
-1.6
-3.2
-1.5
3.7 0.6
-0.9
-0.9
-2.6
1.1
-1.8
-0.9
-2.6
1.8
-2.2


Mean maximum temp (oC)b
33.4 33.4 33.3 33.6 32.3 33.9
33.4 34.1 34.6 32.5
32.8 33.5 30.5
33.8
33.4 32.8 33.0
33.4 34.0 32.7
34.0


Extreme maximum temp (*C)
36.5
36.1 35.5
36.0 35.3
36.6 36.5
36.8 37.2
36.0 35.8 35.9
36.7 36.1 36.0 35.8 35.5
35.4
36.4 35.5
36.7





31


The annual variance of the following monthly values were calculated: mean temperature, mean minimum temperature, mean maximum temperature, extreme minimum temperature, extreme maximum temperature and precipitation. The mean values of these variables for the years 1980-1996 were used for analysis. Host Fly and Host Plant Data


A. suspensa catch data from McPhail traps were obtained from the North Florida Research and Education Center, Quincy, Florida. Trapping was performed by the U.S. Department of Agriculture and by the Florida Department of Agriculture and Consumer Services, Division of Plant Industry. Only trapping data from urban locations were used, in order to conform with parasitoid data, which were also from urban sites. Data were available in the form of monthly numbers of flies per trap for various counties, and exact identity of the town(s) where traps were situated was unknown. Following are the counties for which data were available, and in parentheses the fruit collection site to which they were related in subsequent analyses: Brevard (Melbourne), Broward (Ft. Lauderdale), Charlotte (Punta Gorda), Collier (Naples), Dade (Miami), De Soto (Arcadia), Hardee (Wauchula), Highlands (Lake Placid), Hillsborough (Tampa), Lee (Ft. Myers), Manatee (Bradenton), Okeechobee (Okeechobee), Palm Beach (Belle Glade), Pinellas (St. Petersburg), Polk (Lakeland), Sarasota (Venice), St. Lucie (Ft. Pierce). Trapping data were for the years 1992-1996, except for Brevard, Broward, Dade, Palm Beach and Pinellas Counties, for which data were for 1990-1996. Variables used for analysis included mean monthly catch, minimal monthly catch, and maximal monthly catch. These were calculated for each year, and the mean annual value for each variable (Table 3-3) used for analysis.





32


Table 3-3. Numbers of Caribbean fruit flies captured in McPhail traps for various counties in central and southern Florida. Traps were maintained by the Florida Division of Plant Industry and the U. S. Department of Agriculture. Data were obtained from the North Florida Research and Education Center, Quincy, Florida.


County (corresponding site)
Brevard (Melbourne) Broward (Ft. Lauderdale) Charlotte (Punta Gorda) Collier (Naples) Dade (Miami) De Soto (Arcadia) Hardee (Wauchula) Highlands (Lake Placid) Hillsborough (Tampa) Lee (Ft. Myers) Manatee (Bradenton) Okeechobee (Okeechobee) Palm Beach (Belle Glade) Pinellas (St. Petersburg) Polk (Lakeland) Sarasota (Venice) St. Lucie (Ft. Pierce)


Mean values for the years 1990-1996 for Brevard, Broward, Dade, Palm Beach and Pinellas counties, and 1992-1996 for other counties. aMean monthly catch. bMinimum monthly catch. 'Maximum monthly catch.


Meana
1.4 24 46 35.2 26.5
5.1
6.9
14.8
5.5
34.4
3.0 1.1 8.8
4.0 11.8 19.7
13.4


Minimum
0
0.81 0.51
0.9
1.07 0.55
0.16 0.66 0.36
2.21 0.15
0.02 0.14 0.03 0.72
1.22 1.09


Maximurnc
7.8 115.5
276.7
104.8 129.4 19.0 31.0 55.6
28.0 147.7 15.5
6.1
46.2 22.6 52.3
97.2
42.9





33


A survey was performed to determine the density of host fruit trees in various towns. Quadrants were chosen from various regions of each town. Previous experience suggested that older middle-class neighborhoods were the best areas for A. suspensa hosts. Specific quadrants were picked that appeared on a map to match this designation. Upon arrival, obviously unsuitable quadrants were dismissed, and others chosen from the map to replace them. Four quadrants were sampled in each town, except LaBelle where 5 quadrants were sampled. Host trees were counted during a slow drive through the neighborhood. Thus trees in back yards were counted only if observed from the street. As towns differ in the size and visibility of backyard properties, the number of trees present but not observed per unit area would presumably also differ. Therefore, comparisons of towns based on number of trees counted per unit area may not be reliable. Rather, relative density was estimated as the number of trees observed per km of road. Distances traveled per quadrant ranged from 4.3-14.8 km, but were usually between 5-10 km. Statistical Analysis


Parasitism data were subjected to an arcsine square root transformation before analysis. All analyses were performed using SAS statistical software.

Environmental factors could be associated with either (1) absolute parasitoid distribution, i.e., presence or absence at various sites, or (2) relative abundance of parasitoids among sites in which they are present, as measured by parasitism level. The factors associated with each response may differ.

Associations of environmental factors with presence or absence of each parasitoid species were analyzed using logistic regression models (SAS Institute 1989). All





34


temperature, precipitation and fruit tree density factors were tested separately in these analyses.

Associations of temperature, precipitation and host fruit tree density with parasitism levels were analyzed using linear regression models. Models examining parasitism of all species combined included all sites sampled. Models examining parasitism levels of each species separately included only sites in which the relevant parasitoid species was collected. The various fruit types were analyzed separately. Initially, all temperature, precipitation and fruit tree density factors were included, and their relative fit with the model determined by the forward selection procedure (SAS Institute 1989). Ultimately, only the factor best explaining parasitoid abundance remained in the final single linear regression model. In addition to single regression, multiple regression models were examined including all factors significantly related with parasitism levels.

Host fly population levels are not independent of the previously described factors, i.e., temperature, precipitation and host fruit tree density. Therefore, host fly data could not be included as factors in the previous analyses. Separate linear regression models were examined relating parasitism levels with fly trapping variables. All sites were included in this analysis. Thus in this case I did not differentiate between presence or absence of parasitoid species and their relative abundance.





35


Results


Distribution and Abundance of Parasitoids


Numbers of samples containing parasitoids and numbers of parasitoids emerging for various towns and host fruits are detailed in the Appendix and summarized in Tables 3-4 and 3-5.

With data from all towns combined, parasitism levels were higher in Surinam cherry and Cattley guava than in loquat or common guava for both D. areolatus and D. longicaudata. For U anastrephae, parasitism levels were higher in Surinam cherry than in any other fruit (Table 3-6). Parasitism by U. anastrephae was extremely low in common guava, with only 8 individuals recovered from 4 samples (Table 3-5). These results are consistent with the findings of Sivinski (1991; Sivinski et al. 1997) that smaller fruits have higher levels of parasitism. As expected, the differences in parasitism levels among fruit types is largest for U anastrephae, which has a relatively short ovipositor, and thus less access to larvae deep within large fruits.

Overall, D. areolatus was more abundant than D. longicaudata in Surinam cherry and common guava, but not in loquat or Cattley guava (Table 3-6). Both species were more common than U. anastrephae in loquat and common guava, but mean parasitism levels of D. longicaudata and U anastrephae were not significantly different in Surinam cherry or Cattley guava.

With all data combined, it appears that parasitism levels are quite low (Table 3-6). Note, however, that this includes data from sites where certain parasitoid species were totally absent from samples. Even where parasitoids were recovered, many samples did





36


Table 3-4. Numbers of samples collected and insects emerging for various sites. Each sample includes fruits from a single host tree.

Site Number of samples Number of insects emerging
Total With With With With CFF Dab Dc Uad CFFa Dab Di Uad


Arcadia Belle Glade Bradenton Dade City Ft. Lauderdale Ft. Myers Ft. Pierce Haines City LaBelle Lakeland Lake Placid Lake Wales Melbourne Miami Naples Okeechobee Punta Gorda St. Cloud St. Petersburg Tampa Venice Wauchula


59
71
64 35
54 65
17 31
102 81 70 67
84 48 76
46 54 49 50
37 35 79


W. Palm Beach 25


3095 589
3462 0 1816 0 672 0 2746 0 3092 27 1352 0 281 1 3267 768 3629 172 2837 356 2500 230 961 0 2539 0 1210 5 1748 181 1479 19 1288 0 1354 0 1198 2 2097 1 4368 599 439 0


aCFF = Caribbean fruit fly. "Da Doryctobracon areolatus. 'D= Diachasmimorpha longicaudata. Ua = Utetes anastrephae.


0 106
6
0
214 156
10
0
482
0
0
0
0 81 58
9
1
0
0
0
0
0 51


3 27 33
0 59 79
110
0
7
2
1
0
0 10 36
7 11 16
0
0 60 16
5





37


Table 3-5. Numbers of samples collected and parasitoids emerging for various host fruits. Each sample includes fruits from a single host tree.

Host Number of samples Number of insects emerging
Total With With With With CFFa Dab Dc Uad
CFFa Dab DiC Uad
Loquat 549 418 56 30 13 13618 657 418 63
Surinam cherry 377 326 64 55 63 12329 1197 410 377
Cattley guava 83 60 14 15 7 2147 280 97 34
Common guava 290 261 52 33 4 20395 816 249 8

aCFF = Caribbean fruit fly.
bDa = Doryctobracon areolatus.
*D1 = Diachasmimorpha longicaudata. dUa = Utetes anastrephae.





Table 3-6. Mean percent parasitism (SE) for various host fruits.

Fruit D. areolatus D. longicaudata U. anastrephae
Loquat 2.0 (0.3) B a 1.5 (0.4) B a 0.2 (0.1) B b
Surinam cherry 7.8 (1.1) A a 3.6 (0.6) A b 3.0(0.6) A b
Cattley guava 7.1 (2.8) A a 5.3 (1.6) A ab 1.1 (0.5) B b Common guava 2.5 (0.4) B a 1.0 (0.3) B b 0.03 (0.02) B c

Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio= 100.





38


not contain parasitoids (Tables 3-4 and 3-5). When maximum parasitism is considered, it becomes apparent that all three parasitoid species are capable of achieving high levels of parasitism. Over 50% parasitism was observed in certain samples at 6 sites for D. areolatus, at 5 sites for D. longicaudata, and at 2 sites for U anastrephae (Table 3-7). Additionally, as mentioned above, these observations are almost certainly underestimates of the true parasitism levels.

Mean levels of parasitism for the braconid parasitoids at various sites are summarized in Tables 3-8 through 3-11. At three northern locations, Dade City, Melbourne and St. Petersburg, no parasitoids were found (Table 3-4).

D. areolatus was absent from the Atlantic coast, and also was not collected at Belle Glade, St. Cloud or Bradenton (Figure 3-2). It was most common at interior locations, and relatively rare along the Gulf coast (Figure 3-3). However, distance from the coast was not a significant predictor of D. areolatus abundance, perhaps because of its absence at two interior locations. Highest mean parasitism levels observed for various fruits were 6.6% in loquat at Arcadia, 35.8% in Surinam cherry at LaBelle, 79.7% in Cattley guava at Arcadia, and 10.2% in common guava at Arcadia.

D. longicaudata was not collected at interior locations north and west of Lake Okeechobee, or at the most northern locations along both coasts (Figure 3-2). It was uncommon at all locations at the northern end of its distribution, with the exception of LaBelle (Figure 3-4). D. longicaudata abundance was significantly greater at lower latitudes in both Surinam cherry and common guava (F=12.9, p=0.002 and F=6.5, p=0.02, respectively). Highest mean parasitism levels observed for various fruits were










Table 3-7. Maximum parasitism levels (percent parasitism per sample) in various host fruits at various sites. Includes samples from which at least 10 adult insects were recovered.


Site

Arcadia Belle Glade Bradenton Dade City Ft. Lauderdale Ft. Myers Ft. Pierce Haines City LaBelle Lakeland Lake Placid Lake Wales Melbourne Miami Naples Okeechobee Punta Gorda St. Cloud St. Petersburg Tampa Venice Wauchula W. Palm Beach


Loquat
Daa 29.1
0
0
0
0
3.8
0
0 37.7
0 63.2 22.6
0
0
0
0
4.2

0
1.0
0
32.4
0


Surinam cherry Daa Dlb Uac


'Da = Doryctobracon areolatus. 'Dl = Diachasmimorpha longicaudata.


cUa = Utetes anastrephae.


WJ


DIb
0
0
0
0
14.1 75.0
3.6
0
54.3
0
0
0
0
3.6 1.0
0
0

0
0
0
0 35.9


Ua'
0
0
2.2
0
4.7 9.3 19.6
0
0
0
0
0
0
3.6 7.1
0
0

0
0
0.7
0
0


73.5
0
0

0
7.1
0
0
83.3
24.1 90.9 81.1
0
0
6.0 67.6
40.0
0
0

1.1 87.5
0


0 35.3
14.6

58.5 25.0
4.4
0 72.0
0
0
0
0 51.5
31.3
12.2
0
0
0

0
0
20.0


5.6
64.3 35.2

61.9 21.7 30.9
0
21.1
0
1.2
0
0
15.2
7.5 10.0 22.6
34.3
0

40.0 18.6 11.6


Cattley guava Common guava
Daa Dlb Ua Daa DPb
71.9 0 0 32.4 0
--- --- --- 0 41.2
0 0
0 0
0 0 0 0 17.6
3.8 42.3 6.9 10.8 25.7
--- --- --- 0 0.4
--- --- --- --- --51.9 15.2 0 55.7 24.3
10.9 0 0 32.0 0
65.8 0 0 20.9 0
--- --- --- 29.5 0
0 0 0 0 0
0 0 0 0 12.5
6.7 53.3 15.3 0 14.8
--- --- --- 0 3.1
2.3 2.3 0 --- --0 0
0 0
--- --- --- 0.4 0
--- --- --- --- ----- --- --- 55.9 0
0 0 0 0 0


Uac
0
0
0
0
2.9 3.7
0

0
0.8
0
0
0
0
0
0

0
0
0

0
0





40


Table 3-8. Mean percent parasitism (SE) in loquat for various sites.


Site
Arcadia Belle Glade Bradenton Dade City Ft. Lauderdale Ft. Myers Ft. Pierce Haines City LaBelle Lakeland Lake Placid Lake Wales Melbourne Miami Naples Okeechobee Punta Gorda St. Cloud St. Petersburg Tampa Venice Wauchula W. Palm Beach


D. areolatus
6.6(1.8)
0
0
0
0
0.2(0.1)
0
0
4.9(1.5)
0
3.1 (1.9)
2.7(1.4)
0
0
0
0
0.3 (0.3)
0
0
0.05 (0.05)
0
6.5 (1.5)
0


B
3
3
B
3
B
B
3

3
3

3 AB





3
3
3
3
B








3 B3 3B B3

3B


D. longicaudata
a 0 Cl
0 C
0 C
0 C
2.0 (2.0) BC b 6.9(1.1) AB
0.7 (0.6) BC
0 C
b 10.4(0.3) A
0 C
a 0 C
a 0 C
0 C
1.8(1.8) BC
0.05 (0.05) C 0 C
0 C
0 C
0 C
0 C
0 C
a 0 C
7.2 (7.2) AB


C


Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio= 100.


U. anastrephae
b 0 C b
0 C
0.1 (0.1) C 0 C
0.7(0.7) C
a 1.1 (0.5) BC ab
4.6(3.3) A 0 C
a 0 C c
0 C
b 0 C b
b 0 C b
0 C
1.8(1.8) B 0.4 (0.4) C 0 C
0 C
0 C
0 C
0 C
0.06 (0.06) C
b 0 C b


0





41


Table 3-9. Mean percent parasitism (SE) in Surinam cherry for various sites.


Site
Arcadia Belle Glade Bradenton Ft. Lauderdale Ft. Myers Ft. Pierce Haines City LaBelle Lakeland Lake Placid Lake Wales Melbourne Miami Naples Okeechobee Punta Gorda St. Cloud St. Petersburg Venice Wauchula W. Palm Beach


D. longicaudat


D. areolatus
25.1 (11.1) ABC a 0 E b
0 E
0 E a
0.6 (0.6) E b 0 E b
12.5 (12.5) CDE 35.8 (6.3) A a 4.6 (2.2) E a 20.2 (6.9) BCD a 10.1 (5.5) DE a
0 E
0 E b
0.4 (0.4) E 10.9(5.1) DE a
3.5 (2.8) E 0 E
0 E
0.06 (0.06) E b 32.4 (8.9) AB a
0 E


0 12.1 (3.5)
0.6 (0.6) 7.3 (3.2)
3.6 (2.4) 0.7 (0.4)
0 13.2 (3.5)
0
0
0
0
15.4 (7.2)
2.7 (2.1) 0.8 (0.7)
0
0
0
0
0
3.5 (2.8)


B
A
B AB
B
B
B
A
B
B
B
B
A
B
B
B
B
B
B
B
B


Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio= 100.


a U anastrephae
b 0.4(0.3) A b
a 7.9 (3.3) A a
2.4(1.6) A a 5.1 (2.7) A a
ab 8.4(2.8) A a ab 6.8 (3.4) A a
0 A
b 1.4(1.1) A c
b 0 A b
b 0.07 (0.07) A b
b 0 A b
0 A
a 1.8(1.5) A b
7.3 (6.6) A b 0.6 (0.6) A b
2.4(1.7) A 3.8(2.7) A 0 A
b 4.6(2.0) A a
b 2.0(1.3) A b
1.7(1.7) A





42


Table 3-10. Mean percent parasitism (SE) in Cattley guava for various sites.

Site D. areolatus D. longicaudata U anastrephae
Arcadia 79.7 (10.2) A 0 0 A
Ft. Lauderdale 0 D 0 0 A
Ft. Myers 0.5(0.3) D b 8.8 (3.4) a 1.2(0.8) A b
LaBelle 17.9 (11.9) C 5.7(3.6) 0 A
Lakeland 3.6(3.6) D 0 0 A
Lake Placid 65.8 B 0 0 A
Melbourne 0 D 0 0 A
Miami 0 D 0 10.0 (10.0) A
Naples 0.6(0.6) D 9.8(5.3) 1.8 (1.4) A
Punta Gorda 0.3 (0.3) D 0.3 (0.3) 0 A
W. Palm Beach 0 D 0 0 A

Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio= 100.






10.4% in loquat at LaBelle, 15.4% in Surinam cherry at Miami, 9.8% in Cattley guava at Naples, and 9.6% in common guava at Ft. Myers.

U. anastrephae was widespread, having been collected at most locations (Table 34). Parasitism levels for this species were relatively low, especially at most interior locations (Figure 3-5). Highest mean parasitism levels observed for various fruits were 4.6% in loquat at Ft. Pierce, 8.4% in Surinam cherry at Ft. Myers, 10.0% in Cattley guava at Miami, and 1.0% in common guava at Ft. Myers.





43


Table 3-11. Mean percent parasitism (SE) in common guava for various sites.


Site
Arcadia Belle Glade Bradenton Dade City Ft. Lauderdale Ft. Myers Ft. Pierce LaBelle Lakeland Lake Placid Lake Wales Melbourne Miami Naples Okeechobee St. Cloud St. Petersburg Tampa Wauchula W. Palm Beach


Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio= 100.


D. areolatus
10.2 (4.1) A 0 B
0 B
0 B
0 B
2.4(2.1) AB
0 B
7.1 (2.8) AB 2.5(1.4) AB 4.9 (2.0) AB 4.5 (1.8) AB
0 B
0 B
0 B
0 B
0 B
0 B
0.05 (0.05) B
6.2 (2.9) AB
0 B


D. longicaudata
a 0 B
2.0(1.4) B 0 B
0 B
b 3.5(1.7) B
9.6 (4.4) A 0.4 B
a 3.0(1.2) B
a 0 B
a 0 B
a 0 B
0 B
b 1.3(0.7) B
4.9 (4.9) AB
0.8 (0.8) B 0 B
0 B
0 B
a 0 B
0 B


U. anastrephae
b 0 B b
0 B
0 B
0 B
a 0.2 (0.2) B b
1.0(0.7) A 0 B
ab 0 B b
b 0.03 (0.03) B a b 0 B b
b 0 B b
0 B
a 0 B b
0 B
0 B
0 B
0 B
0 B
b 0 B b
0 B


b





44


The distribution ranges of D. areolatus and D. longicaudata overlap only within a limited area from Lake Okeechobee to the Gulf of Mexico coast (Figure 3-2). In fact, only at LaBelle are both common (Figure 3-6). Interestingly, LaBelle was among the sites with the highest parasitism levels for both D. areolatus and D. longicaudata. With all sites included, there was no significant relationship between parasitism levels of D. areolatus and D. longicaudata in Surinam cherry (Spearman correlation coefficient =

-0.27, p= 0.23). However, with LaBelle excluded, parasitism by the two species was negatively related (Spearman correlation coefficient = -0.48, p=0.034). Temporal and spatial dynamics at LaBelle are explored in Chapter 4.

Although U anastrephae is widespread, it is most common at coastal locations and at Belle Glade, i.e., at locations where D. areolatus is absent or rare and where D. longicaudata is common. With all sites included, parasitism levels of U anastrephae and D. longicaudata in Surinam cherry were positively related (Spearman correlation coefficient = 0.53, p=0.014), indicating a similar distribution pattern for both species. Interestingly, when considering only sites where both species were recovered, no significant relationship was observed (Spearman correlation coefficient = -0.02, p=0.96). This suggests that these species do not impact each other on a local level.

With all sites included, no significant relationship was observed between parasitism levels of U anastrephae and D. areolatus in Surinam cherry (Spearman correlation coefficient = -0.35, p=0.12). However, when only sites with both species present were considered, parasitism was negatively related (Spearman correlation coefficient = -0.70, p=0.044). This suggests that U anastrephae and D. areolatus may have a negative impact on each other. Only at Punta Gorda are these species found





45


together in similar numbers (Figure 3-6). Note that at this location parasitism by both species was low, with only 19 D. areolatus and 11 U anastrephae recovered (Table 3-4). Mean parasitism at Punta Gorda in Surinam cherry was only 3.5% and 2.4% for D. areolatus and U anastrephae, respectively (Table 3-9). Thus significant competition at this site was unlikely.

A fourth parasitoid, the eulophid Aceratoneuromyia indica Silvestri, emerged from two puparia collected from guava at Belle Glade in February 1995. To my knowledge, this is the only report of this species outside the area of its introduction in Dade County, ca. 140 km to the south.


Host Plant Density


Sample sites varied considerably in the relative density of host fruit trees (Table 3-12). Loquats are most abundant at Dade City, Lake Wales, Melbourne and Tampa, all of which are situated at high latitudes (Figure 3-1). At Dade City, the highest latitude site in this study, other hosts are rare, with only one small Surinam cherry hedge and one common guava located. Lowest numbers of loquats were found at Ft. Myers, Miami, Ft. Lauderdale and Belle Glade. Surinam cherries are most abundant at Miami, St. Petersburg and Belle Glade, Cattley guavas at Naples, Ft. Pierce and LaBelle, and common guavas at LaBelle. Note that at southern coastal towns like Miami, additional tropical host trees occur, which are not included in this survey. The most important of these hosts is tropical almond, Terminalia catappa L.




46


Limit of Doryctobracon areolatus
distribution

0
Atlantic
0 Ocean
0
0
0

Gulf of
Mexico e
so
Northern limit of Diachasmimorpha longicaudata
distribution

o Doryctobracon areolatus
* Diachasmimorpha longicaudata




Figure 3-2. Distribution of Doryctobracon areolatus and Diachasmimorpha longicaudata. Includes data from the current study and Sivinski et al. (1996).





Figure 3-3. Parasitism by Doryctobracon areolatus in Surinam cherry.


47


o Atlanti
Ocean







f of ico 0 0


0%
<1% 1-5% 5-10%

10-20%

>20% Release site


Gul Mex


0


6


C





















Gul Me


0 AtIanti~


Figure 3-4. Parasitism by Diachasmimorpha longicaudata in Surinam cherry.


48


o Atlanti 0 0 Ocean
0



0




f of KiCO




0% <1% 1-5% 5-10% S10-20%


0


C



















Gu Me


Figure 3-5. Parasitism by Utetes anastrephae in Surinam cherry.


49


Atlantic
0 0 Ocean







If of xico



DO 0% 0 <1% j1-5% 5-10%











L


Figure 3-6. Relative abundance of parasitoid species in Surinam cherry in the region of co-occurrence. Da= Doryctobracon areolatus; Dl= Diachasmimorpha longicaudata; Ua= Utetes anastrephae. The solid line represents the northern limit of distribution for D. longicaudata.


tA


radenton Wauchula
Lake Ok chobee
Placid Arcadia Venice L k


Okeechobee

Gulf of Puonta Mexico Ft. ers -LaBelle Belle Glade


Ua
Da

~~DI





51


Table 3-12. Mean host fruit plant density (trees/ km of road) (SE) in various towns.


Town Arcadia Belle Glade Bradenton Dade City Ft. Lauderdale Ft. Myers Ft. Pierce LaBelle Lakeland Lake Placid Lake Wales Melbourne Miami Naples Okeechobee Punta Gorda St. Cloud St. Petersburg Tampa Venice Wauchula


Means within a column followed by the same letter are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio=100.


Loquat
3.4 (0.4) cde
2.2 (0.5) ef
3.3 (0.6) cde
6.5 (0.4) a 1.8 (0.4) ef 0.9 (0.2) f
4.4 (0.5) abcd 3.0 (0.8) cdef 3.5 (1.4) ede 2.9 (1.2) cdef
6.1 (0.6) ab 5.9 (1.1) ab
1.1 (0.4) f
2.5 (0.3) def 3.1 (0.2) cdef 3.9 (1.7) bcde
3.4 (0.6) cde
3.0 (0.2) cdef
5.8 (1.1) ab
2.4 (0.3) def 4.8 (0.6) abc


Surinam cherry
2.6 (0.4) ef 7.7 (0.8) bc
4.6 (0.7) cde 0.04 (0.04) f 5.3 (0.9) bcde
2.7 (0.3) ef
4.9 (0.9) cde
3.4 (0.8) e 3.7 (0.4) e
4.9 (0.9) cde 7.4 (0.8) bcd
3.2 (1.0) ef 12.9 (3.9) a 4.8 (0.3) cde
3.7 (0.4) e
4.2 (1.1) de 2.3 (0.6) ef 8.2 (2.0) b 2.6 (0.5) ef 2.3 (0.6) ef 3.5 (1.0) e


Cattley guava
0.8 (0.1) bcd 0.2 (0.02) cd
0.3 (0.1) cd
0 d
0.5 (0.2) bcd 0.6 (0.1) bcd
1.4 (0.7) b 1.0 (0.2) bc
0 d
0.5 (0.2) bcd 0.5 (0.2) bcd
0.3 (0.1) cd 0.3 (0.2) cd
3.9 (1.2) a
0.5 (0.1) cd 0.3 (0.1) cd 0.2 (0.1) cd 0.2 (0.1) cd 0.2 (0.1) cd
0.6 (0.2) bed 0.8 (0.2) bed


Common guava
0.10 (0.06) b
0.5 (0.3) b
0.02 (0.02) b 0.04 (0.04) b
0 b
0.06 (0.04) b
0.2 (0.1) b 2.4 (0.9) a
0.06 (0.04) b
0.5 (0.3) b 0.4 (0.2) b 0.3 (0.1) b 0.7 (0.3) b
0.10 (0.06) b
0.6 (0.3) b
0 b
0.2 (0.1) b
0.06 (0.03) b 0.04 (0.04) b
0 b
0.8 (0.2) b





52


Relationships With Environmental Factors

The northern limit of D. longicaudata distribution closely fits the isotherm of 10.5* C January mean minimum temperature (Figure 3-7). The number of frost days also fits the distribution somewhat (see map in Fernald 1981), but to a lesser degree. Mean minimum temperature for the coldest month was not a significant factor in the logistic regression analysis (X2=3.51, p=0.061). Rather, presence of this parasitoid was best explained by low variability in monthly temperatures (Table 3-13). Four different variance factors were negatively related with presence of D. longicaudata, including variance of extreme maximum, extreme minimum, mean maximum and mean temperature. Other factors significantly related with presence of D. longicaudata included mean and extreme minimum temperatures (positive relationships) and abundance of loquat trees (negative relationship). In contrast with D. longicaudata, variance of temperatures had a positive relationship with presence of D. areolatus. In this case the significant factor associated with parasitoid distribution was variance of extreme minimum temperatures (Table 3-13). Presence of D. areolatus was also positively associated with extreme maximum temperatures. Note that summer temperatures are often greater at northern and inland sites than at southern and coastal locations (Table 3-2). Presence of U. anastrephae showed no significant relationships with abiotic factors, but was negatively related with abundance of loquat trees.

Density of guava trees was positively related with total parasitism in both loquat and Surinam cherry (Table 3-14). Other factors significantly associated with total parasitism include variance of extreme minimum temperatures (positive relationship in Cattley guava), extreme maximum temperature (positive relationship in common guava)


















Gu Me


(


Figure 3-7. Relationship between the northern observed limit of Diachasmimorpha longicaudata distribution and January mean minimum temperature. The isotherm was copied from Fernald (1981), based on data from the years 1960-1979.


53


AtIanti
Ocean







If of
xico


Northern limit of distribution


10.50C January
average minimum
temperature


c





54


and variance of precipitation (positive relationships in loquat and common guava, multiple factor models).

D. areolatus was more common at sites with lower mean temperatures, in both Surinam cherry and Cattley guava (multiple and single factor models, respectively; Table 3-15). Similarly, parasitism levels in common guava were negatively associated with extreme minimum temperatures (multiple factor model). Variance of extreme maximum temperatures was negatively related with D. longicaudata abundance in Surinam cherry (multiple factor model). Variance of precipitation was positively associated with D. areolatus abundance in Cattley and common guava, and with D. longicaudata abundance in loquat (multiple factor models).

Guava tree density was significantly associated with parasitism levels of both D. areolatus (positive relationship in Surinam cherry) and D. longicaudata (positive relationships in loquat and Surinam cherry, multiple factor models) (Table 3-15). Loquat tree density was negatively related with D. longicaudata parasitism levels in Surinam cherry and common guava (multiple factor models). Similarly, Surinam cherry tree density was negatively related with D. longicaudata abundance in common guava (multiple factor model). A highly significant positive relationship was observed between Surinam cherry tree density and parasitism levels of U. anastrephae in Cattley guava. Interestingly, there were no significant relationships between tree density and parasitism on the same host.

Highly significant positive relationships were found between fly trapping and parasitism by D. longicaudata in loquat and common guava (minimum and mean monthly capture of host flies, respectively; Table 3-16). Similar, but less significant,





55


Table 3-13. Environmental factors significantly associated with presence or absence of parasitoid species, according to logistic regression analysis.

Parasitoid species Factor (relationship) p X
D. areolatus Var extreme min temp (+) 0.039 4.26
Extreme max temp (+) 0.045 4.00
D. longicaudata Var extreme max temp (-) 0.016 5.79
Var extreme min temp (-) 0.017 5.69 Var mean max temp (-) 0.028 4.82 Var mean temp (-) 0.030 4.70
Mean temp (+) 0.035 4.46
Extreme min temp (+) 0.038 4.28
Loquat density (-) 0.039 4.25
U. anastrephae Loquat density (-) 0.016 5.78

Extreme min temp = Extreme minimum annual temperature. Extreme max temp = Extreme maximum annual temperature. Loquat density = Mean density of loquat trees. Var extreme min temp = Variance of monthly extreme minimum temperatures. Var extreme max temp = Variance of monthly extreme maximum temperatures. Var mean temp = Variance of monthly mean temperatures. Var mean max temp = Variance of monthly mean maximum temperatures.




relationships were observed between minimum capture and abundance of U. anastrephae in loquat and Surinam cherry. No significant relationships were found between fly trapping variables and parasitism by D. areolatus.


Discussion


Factors possibly affecting parasitoid distribution and abundance can be grouped into three categories: (1) abiotic factors, e.g. temperature and precipitation; (2) host





56


Table 3-14. Environmental factors significantly associated with parasitism levels of all parasitoid species combined.


Host fruit


Single factor model
Factor (relationship) pC r2


Multiple factor modelP
Factors (relationship) pc r2


Loquat Guava density (+) ** 0.34 Guava density (+) * 0.61
Var precipitation (+) **
Surinam cherry Guava density (+) 0.52 --Cattley guava Var extreme min temp * 0.36 ----(+)


Common guava Extreme max temp (+) * 0.21 Extreme max temp (+) * 0.42 Var precipitation (+) *

aLinear regression with the factor which best explains variation in the data. bMultiple linear regression with factors having significant linear relationships with the relevant parasitism level.
c* p<0.05; ** p<0.01; *** p<0.001. Guava density = Mean density of common guava trees. Extreme max temp = Extreme annual maximum temperature. Var extreme min temp = Variance of the monthly extreme minimum temperatures. Var precipitation = Variance of the monthly precipitation.


availability; and (3) competition with other species. These environmental effects are summarized in Figure 3-8.

The northern limit of D. longicaudata distribution is closely related to the January mean minimum isotherm of 10.5*C, as derived from Fernald (1981). This suggests that winter temperatures limit the distribution of this species. This could be an important factor affecting tropical parasitoid species in warm temperate localities. Snowball and Lukin (1964) suggested that winter temperatures may limit the establishment of Fopius arisanus in Australia.

Sivinski et al. (1998) reported a reduction in winter in the abundance of D. longicaudata relative to that of D. areolatus in calamundin at LaBelle. This suggests





57


Table 3-15. Environmental factors significantly associated with parasitism levels of the various parasitoid species, for towns in which the relevant parasitoid was collected.


Species Host fruit Single factor models Multiple factor models
Factor (relationship) pd r2 Factors (relationship) pd r2 Da Loquat ----- -----


Surinam cherry Cattley guava


Guava density (+) Mean temp (-)


Common guava


Dl Loquat


Surinam cherry


Cattley guava Common guava


* 0.50 Guava density (+)
Mean temp (-)
** 0.90 Mean temp (-)
Var precipitation (+) Extreme min temp (-) Var precipitation (+)


Guava density (+) Var precipitation (+) Var extreme max temp (-)
Guava density (+) Loquat density (-)


Loquat density (-) Sur cherry density (-)


**
**

*


0.81 0.98


0.84
**


* 0.68
*

** 0.87

***


*
*


0.75


Ua Loquat
Surinam
cherry
Cattley guava


Sur cherry density (+)


'Da = Doryctobracon areolatus; D 1= Diachasmimorpha longicaudata; Ua = Utetes anastrephae. bLinear regression with the factor which best explains variation in the data. cMultiple linear regression with factors having significant linear relationships with the relevant parasitism level.
d* p<0.05; ** p<0.01; *** p<0.001. Sur cherry density = Mean density of Surinam cherry trees. Guava density = Mean density of common guava trees. Loquat density = Mean density of loquat trees. Extreme min temp = Extreme annual minimum temperature. Mean temp = Mean annual temperature. Var extreme max temp = Variance of the monthly extreme maximum temperatures. Var precipitation = Variance of the monthly precipitation


** 0.68





58


Table 3-16. Fly trapping variables significantly associated with mean parasitism levels. Parasitoid Host fruit Variable (relationship)' pb r2
D. longicaudata Loquat Minimum (+) *** 0.64
Common guava Mean (+) *** 0.68 U. anastrephae Loquat Minimum (+) ** 0.38
Surinam cherry Minimum (+) * 0.30 aLinear regression with the factor which best explains variation in the data. Mean = Mean monthly capture (flies/trap). Minimum = Minimum monthly capture (flies/trap). b* p<0.05; ** p<0.01; *** p<0.001.



some climatic effect unrelated to host availability. A similar conclusion was reached by Snowball (1966) regarding abundance of F. arisanus in southern Australia. Laboratory studies suggest that D. longicaudata may be less tolerant to low temperatures than its host A. suspensa (Chapter 5). Further studies are needed to determine whether its tolerance is lower than that of D. areolatus.

Note that mean minimum temperature was not a significant factor in the logistic regression analysis. Rather, absence of D. longicaudata was best explained by high variability in temperatures (Table 3-13). The possible significance of this observation is discussed below.

Parasitoid populations interact with the temporal and spatial distribution of their hosts, and the dynamics of both host fruits and fly populations may be important. The three towns from which parasitoids were not collected all have some type of low host availability. At Dade City, the most northern town in this study, all host plants except loquat are absent or rare (Table 3-12). Thus, hosts are available for only a short period of time in early spring. Note that host larvae are not uncommon in loquat at Dade City, with





59


91% of samples collected producing flies (Table 3-4). Host fly densities in Brevard county (Melbourne) are among the lowest of all counties included in this study (Table 33). Only 63% of samples collected at Melbourne produced flies (Table 3-4). Hosts are also relatively rare in Pinellas county (St. Petersburg) (Table 3-3), with only 68% of samples collected producing flies (Table 3-4). Note also that Pinellas county is a peninsula, separated form the east and south by Tampa Bay. This body of water may be an ecological barrier to parasitoids, making colonization difficult.

Fluctuations in fruit availability would depend primarily on abiotic conditions including temperature and precipitation (Petr 1991, Raper and Kramer 1983). High variability in temperatures could lead to greater heterogeneity in the temporal occurrence of fruit. Thus the negative relationships between the variances of several temperature variables and occurrence of D. longicaudata (Table 3-13) may indicate that this parasitoid is dependent on a relatively constant supply of hosts. This hypothesis is supported by the highly significant relationship between numbers of flies trapped and parasitism by this species in loquat and guava (Table 3-16).

Spatial distribution of host fruits is dependent upon (1) the number of fruits per tree and (2) the number of trees per unit area. The number of fruits per tree depends on various factors including tree size and age, degree of shading, and horticultural practices such as pruning, watering and fertilization. Although these factors may vary among towns, no data on this are available.

In several cases, parasitism appears to be related to host tree density. Guava tree density accounts for 34 and 52% of the variation in total parasitism for loquat and Surinam cherry, respectively (Table 3-14). It is a significant factor in the abundance of





60


both D. areolatus and D. longicaudata (Table 3-15). Guava is usually the last available major host before the onset of winter. Thus its abundance may be an important determinant of the size of the overwintering population, which in turn affects the abundance of parasitoids in loquat and Surinam cherry the following spring.

In Hawaii, D. longicaudata is more common in orchards than in wild guava (Vargas et al. 1993). The authors suggest that two possible factors contributing to this observation may be high tree densities and abundance of rotting fruit in commercial guava orchards.

The density of Surinam cherry trees accounts for 68% of the variation in U. anastrephae abundance in Cattley guava (Table 3-15). This result is expected because Surinam cherry is the major host for U anastrephae (Table 3-6), and it immediately precedes Cattley guava in fruiting. Note, however, that U anastrephae was recovered from Cattley guava in three towns only (Table 3-10), and a regression analysis based on three points should be treated with caution.

Significant negative relationships between fruit tree density and parasitism are probably artifacts. The absence of U anastrephae in towns with high incidence of loquats may reflect a tendency of loquats to survive in towns that have an unsuitable climate for other tropical fruits, and that the lack of the latter is what actually accounts for the absence of parasitoids (loquats flourish and fruit well north of the normal range of A. suspensa, pers. obs.). Dade City, Melbourne and Tampa, among the towns with the highest loquat densities, all have relatively low densities of other host fruits (Table 3-12).

When all towns except LaBelle are considered, abundance of D. areolatus is negatively related to that D. longicaudata (Figures 3-3, 3-4 and 3-6). Diachasmimorpha





61


longicaudata and D. areolatus are of similar size, and both have long ovipositors (Sivinski et al. 1997). They also show similar preferences to the host fruits (Table 3-6), and attack the same stage larvae (Lawrence et al. 1976, Chapter 6). Thus the potential for competition between these species is obvious. Note that the interaction between them is a new association, as D. areolatus is a neotropical species, while D. longicaudata originates in the Indo-Pacific region (Clausen 1978). Therefore, they would not have evolved niche divergence to avoid competition.

D. areolatus was established in large numbers after being introduced to southern Florida (Baranowski and Swanson 1970), but has subsequently diminished and possibly disappeared from the region of its introduction. The current distribution pattern suggests that competition by D. longicaudata may have caused its displacement. Note that D. longicaudata was introduced three years following the introduction of D. areolatus. Thus D. areolatus had time to become established and migrate to favorable locations to the north prior to the establishment of D. longicaudata. It is intriguing to think that had the sequence of introductions been reversed, D. areolatus may not have successfully been established.

Similar cases involving fruit fly parasitoids in Hawaii are considered among the classic examples of apparent competitive displacement. Psyttalia (Opius) humilus (Silvestri) was the dominant parasitoid of the C. capitata in 1915, and was replaced by Diachasmimorpha tryoni (Cameron) from 1916 onward (Pemberton and Willard 1918). This displacement was apparently due to competitive superiority of the first-instar larvae of the latter species. During the late 1940s, several parasitoid species were released for the control of B. dorsalis (Clausen et al. 1965). Initially, D. longicaudata was the





62


dominant species, only to be replaced by Fopius (Biosteres) vandenboschi (Fullaway), which was in turn replaced by Fopius (Biosteres) arisanus (Sonan) (van den Bosch et al. 1951). Two contributing mechanisms were offered for this displacement. First, the displacing species attack progressively earlier immature stages, which would be more prone to parasitism because of their proximity to the fruit surface. Second, F. arisanus larvae appeared to inhibit the development of the other species, while F. vandenboschi larvae also inhibit development of D. longicaudata (van den Bosch and Haramoto 1953).

In contrast to the report of van den Bosch and Haramoto (1953), several studies indicate that D. longicaudata may have a competitive advantage over other parasitoid species in situations of multiparasitism. Palacio et al. (1991) found that D. longicaudata was a superior competitor to both F. arisanus and Fopius (Biosteres) persulcatus Silvestri, indicating physical competition among first-instar larvae. Ramadan et al. (1984) suggested a similar advantage of D. longicaudata over D. tryoni in C. capitata hosts. Studies by Bautista and Harris (1997) with D. longicaudata and Psyttalia incisi (Silvestri) indicate that the sequence of oviposition is important, with the first parasitoid species to which the host is exposed having an advantage. However, while exposure first to P. incisi resulted in 77% of the progeny being of this species, the reverse sequence resulted in 99% of the progeny being D. longicaudata. These studies support the possibility that D. longicaudata may be a superior competitor to D. areolatus in multiparasitized hosts.

To summarize, the two major phenomena observed in this study are the absence of D. longicaudata in the interior region of central Florida, and the absence of D. areolatus in portions of southern Florida. Which factors may affect the interactions





63


between D. areolatus and D. longicaudata, and how could this result in the observed pattern of distribution? Sivinski et al. (1998) hypothesized that the co-occurrence of both species at LaBelle may be the result of "counter-balanced competition" (cf. Zwlfer 1971) where D. areolatus is superior to D. longicaudata in locating host patches (=extrinsic competitor) and D. longicaudata is superior in exploiting these patches (=intrinsic competitor). The better searcher would be at an advantage at locations that have a less predictable supply of hosts in time or space, while the better intrinsic competitor would benefit from more homogeneous host availability. In the more northern interior regions of Florida, where temperatures are more variable, large gaps may occur between fruiting cycles of the various hosts, and in particular between the fall fruiting of guava and the spring fruiting of loquat. At coastal locations where temperature conditions are more homogeneous, trees may have more than one fruiting cycle, filling in the temporal gaps in fruit availability (Nguyen et al. 1992). Furthermore, additional tropical host fruits occur in the southern coastal regions (see Hennessey 1994). The former conditions would favor the superior searcher, presumably D. areolatus, while the latter would benefit the superior intrinsic competitor, i.e., D. longicaudata. In extreme conditions one parasitoid species may driven to extinction, and at intermediate locations both would persist. At LaBelle large temporal gaps in hosts may be balanced by spatial abundance, in particular of guava, enabling the persistence of a sizable population of D. longicaudata.

Diapause development is an important mechanism allowing insects to cope with periods of low host availability. There is evidence that D. longicaudata individuals do indeed enter diapause (Aluja et al. submitted, Ashley et al. 1976, Clausen et al. 1965,





64


Chapter 5). However, in Mexican populations both the proportion of individuals entering diapause and the length of diapause period are greater for D. areolatus than for D. longicaudata (Aluja et al. submitted). Additionally, Aluja et al. (submitted) report circumstantial evidence that D. areolatus adults may enter a reproductive diapause. These observations suggest that D. areolatus populations may be able to better survive long periods without hosts.

The presumed competitive advantage of D. longicaudata over D. areolatus may be the result of other mechanisms, besides larval competition. Diachasmimorpha longicaudata may have an advantage in locating fruits containing host larvae, or in locating larvae within fruits. Diachasmimorpha longicaudata locates hosts within fruits by sensing the vibrations of the feeding larvae (Lawrence 1981). In the laboratory D. longicaudata females can locate hosts without the presence of host fruit odors. On the other hand, host larvae alone are not attractive to D. areolatus females, and addition of fresh fruit odors is sufficient to stimulate oviposition behavior in this species (Chapter 7). It is possible that D. areolatus may respond to host vibrations following exposure to host fruit odors. However, its dependence on fruit odors suggests that vibrations may be a relatively less important stimulus for D. areolatus than for D. longicaudata, and consequently it may be at a disadvantage in locating larvae within fruit. Conversely, this dependence on fruit odors may indicate a superior ability of D. areolatus to locate host patches. Note, however, that D. longicaudata females are attracted to volatiles associated with rotting fruit (Greany et al. 1997).

Additionally, D. longicaudata would have a competitive advantage if it were more fecund. In the laboratory, D. longicaudata can produce a large number of progeny





65


over a short period of time, with most eggs laid within a few days (Greany et al. 1976). On the other hand, D. areolatus appears to produce smaller numbers of progeny over longer periods of time (Chapter 6). If this is the case in the field, an individual D. longicaudata female, after locating the host patch, could exploit it at a faster rate than a D. areolatus female.

Finally, the ovipositor of D. longicaudata is longer than that of D. areolatus (D. longicaudata, 5.49 0.21 mm, range 4.67-6.40 mm, n=7; D. areolatus, 3.80 0.11 mm, range 3.43-4.11 mm, n=7; t=7.18, p<0.0001). The ovipositor of D. longicaudata is also consistently longer in relationship to body size as estimated by the ratio between ovipositor length and wing length (t=14.59, p<0.0001, n=7). This enables D. longicaudata to reach larvae deeper within fruits, thus allowing access to a larger proportion of larvae, especially in large fruits (Sivinski et al. 1997).

As discussed above, D. longicaudata distribution may be limited to the north by

(1) direct effects of low winter temperatures or (2) periods of low host availability. There is some evidence in support of both hypotheses. The two effects could act in concert, and may not be mutually exclusive.

Some support for the second hypothesis can be found in the literature. Preliminary observations in Mexico, suggest that D. longicaudata is less common at low altitudes (M. Aluja, pers. comm.). This is the opposite of what would be expected if this species was adversely affected by low temperatures. The low altitude habitats are drier, and as a result there are larger gaps in host availability. Thus, the temporal availability of hosts at low altitudes in Mexico is similar to that in the colder regions of Florida.





66


In Mexico, D. longicaudata is the dominant species in an area of mixed cultivation, while it is absent in native habitats where D. areolatus is most common (Aluja et al. 1990, Hernandez-Ortiz et al. 1994). Similarly, in Amazonas State, Brazil, D. areolatus is the dominant parasitoid in rural locations while Opius sp. nr. bellus is dominant in urban areas (Canal D. et al. 1995). Native habitat is more heterogeneous in host availability, favoring a superior searcher, while there is a more predictable supply of hosts in cultivated or urban areas, which would flvor the better intrinsic competitor. In both cases, the presence of wild hosts may give D. areolatus refuge from competition, thus preventing displacement. Such a refuge does not exist in Florida, where most hosts are in either urban or agricultural habitats. Feral populations of guavas exist in some areas, but fruiting of these trees is mostly limited to late summer.

The biology of U. anastrephae is little known, and its interactions with other parasitoid species unclear. Laboratory rearing data from Brazil suggest that it attacks the same larval instars as D. areolatus and D. longicaudata (R. Sugayama, pers. comm.). In Mexico, Sivinski et al. (1997) observed negative relationships between U. anastrephae and D. areolatus within tree canopies. This was interpreted as being the possible result of evolution of divergent niches in these sympatric species, which would reduce direct competition. However, the inverse among-site relationship observed between these two species in this study suggests that significant within-site competition may be occurring. Note that U. anastrephae is common only in small fruits such as Surinam cherry (Table 3-6). Thus significant competition would only occur in such fruits. As D. areolatus was established in Florida with U anastrephae already present, in appears that competition by U. anastrephae alone is not highly significant. However, U. anasirephae occurs in large





67


numbers in most of the same towns where D. longicaudata is common (Figures 3-4 through 3-6). Thus, coupled with the competitive pressure of D. longicaudata, it may have contributed to the displacement of D. areolatus. Alternatively, U. anastrephae may be common at these locations because D. longicaudata had suppressed D. areolatus, thus releasing the former species from competition.

There has been some debate over the merits of multiple introductions of natural enemies in biological control programs. Some workers have suggested that interspecific competition may reduce the overall level of host suppression, while others claim that additional species would increase control levels (see discussion in Van Driesche and Bellows 1996). DeBach and Rosen (1991) state that "displacement of a fairly effective established natural enemy species by another imported species means that the second one is more effective, and will produce even better host population regulation". Could the displacement of D. areolatus by D. longicaudata have reduced the suppression of the Caribbean fruit fly? This scenario appears to be possible. Mean parasitism levels for D. areolatus in Surinam cherry surpassed 20% at four towns, and at LaBelle reached 36% (Table 3-9). Meanwhile, mean parasitism for D. longicaudata was no higher than 15% (at Miami). Even with D. longicaudata and U. anastrephae combined, the highest mean parasitism level was 20% (at Belle Glade). Thus it appears that D. areolatus may be able to contribute to higher levels of control than the other parasitoid species, possibly due to its greater searching efficiency.

If this is the case, how then could this more efficient parasitoid be displaced? We could attempt to understand parasitoid population dynamics on a regional scale by examining the possible dynamics within individual host trees. Consider two Surinam





68


cherry trees. The first produces large numbers of fruits over a short period of time. Host larvae would be abundant, and parasitoids would be limited by the number of fruits and larvae that they could locate. As more individuals of the superior searcher would be able to locate the resource quickly, on a population level it would be capable of locating greater numbers of larvae. The second tree produces fruits over a long period of time in limited patches. The superior searcher would have the initial advantage of colonizing the resource more quickly. However, after the superior intrinsic competitor finally locates the resource, the latter species would have the advantage. While over time the former species may parasitize on average a higher proportion of hosts, towards the end of the fruiting cycle the latter species would dominate and possibly displace the superior searcher. Such dynamics (though not leading to displacement) where observed within trees at LaBelle, with both the relative and absolute numbers of D. longicaudata increasing over time (Sivinski et aL 1998). Note that parasitism by D. longicaudata can approach 100% at the end of the fruiting period in fruits such as Surinam cherry. If the majority of host trees within a town were of the second type, total displacement could occur. Fruiting patterns within trees are dependent on weather conditions, and could change over time. It is conceivable that in certain years most trees would be of the second type, and the population dynamics would lead to displacement. In other years the first type may dominate, and if displacement had previously occurred, the remaining parasitoid would not sufficiently respond to the growing host population, and parasitism levels would be low. In conclusion, a superior intrinsic competitor may displace a superior searcher, leading to a reduction in host suppression.





69


D. areolatus is absent or rare at northern coastal locations, even though total parasitism at these sites is low. It appears that other factors besides competition must account for this absence. This suggests that unidentified environmental factors associated with coastal locations may have contributed to its disappearance from southern Florida as well. However, if coastal conditions were unsuitable to D. areolatus, it would not have been expected to become established in large enough numbers to enable it to spread to distant regions of the state. Perhaps widespread pesticide applications against mosquitoes and other biting insects, which is most prevalent in coastal regions, contributed to its disappearance.

Could D. areolatus have displaced D. longicaudata in central Florida, just as D. longicaudata may have displaced D. areolatus in the south? D. areolatus obviously migrated from its original release area in southern Florida to the areas in which it currently dominates in the central part of the state. However, it is unclear whether it was present there when D. longicaudata was released in 1972, only three years after its own introduction. Could a superior searcher displace a superior intrinsic competitor? Although a superior searcher would have a relative advantage in a situation of less predictable hosts, allowing it to be more successful on a population level, the superior competitor by definition would have an advantage when both species occur together on a patch, regardless of the mechanism of competition. It is more likely that the superior competitor would be driven to extinction due to lack of hosts than because of effects of a less competitive parasitoid species.

Is it advisable to release D. longicaudata in periodic inundative releases in the regions of central Florida from which it is presently absent? It is quite easy to rear D.





70


longicaudata in the laboratory, and mass-rearing can be achieved at relatively low cost. In contrast, D. areolatus is much more difficult to rear (Chapter 6). Therefore, it would be cost-effective if the program of inundative releases of D. longicaudata could be expanded to all regions of Florida (see Burns et al. 1996). There is evidence that largescale inundative releases of D. longicaudata could reduce host populations (Sivinski et al. 1996). The numbers of parasitoids released is presumably much higher than those naturally occurring in the field (Knipling 1992). Thus, there is no real difference between augmenting existing populations and releases in areas where parasitoids do not occur. There is no reason to believe the parasitoids would not be as effective in all regions of the state, at least during warm periods of the year. However, there is a substantial risk that these releases would cause the permanent displacement of D. areolatus. If releases are terminated after such a displacement, D. longicaudata would not be expected to become permanently established (because they would not survive the winter or periods lacking in hosts), and no parasitoids would remain. Thus initiation and subsequent termination of an inundative release program for D. longicaudata could ultimately lead to an explosion of A. suspensa populations. If inundative parasitoid release in central Florida is pursued, it may be more advisable to develop more cost-effective rearing procedures for D. areolatus, with the objective of releasing this species in areas where it currently occurs.








Low temperatures


Variable temperatures


I I (-)


(+)


U (+)


(+)


Surinam cherry tree density

(+)


rephae


(D. areolatus U. anas


Figure 3-8. Summary of factors possibly affecting parasitoid abundance.


(-)


Predictable fruit availability


Guava tree density


D. longicaudata

(-)













CHAPTER 4
LOCAL TEMPORAL AND SPATIAL DISTRIBUTION PATTERNS OF
DIA CHASMIMORPHA LONGICA UDA TA AND DORYCTOBRACON AREOLA TUS IN
AN AREA OF CO-OCCURRENCE


The distributions of Diachasmimorpha longicaudata and Doryctobracon areolatus overlap in Florida only within a limited region (Chapter 3). The town of LaBelle, situated between Lake Okeechobee and the Gulf of Mexico, is one of the few locations in which both are common (Chapter 3, Sivinski et al. 1996, 1998). Studies of the temporal and spatial dynamics of these parasitoids at LaBelle may help explain how they co-occur at this location, while in most areas of Florida they do not. More specifically, they could address hypotheses generated in Chapter 3, i.e. that low temperatures have a direct or indirect negative effect on D. longicaudata, and that interspecific competition could partially explain the disappearance of D. areolatus from southeastern Florida.

Temporal and spatial dynamics of these species within trees were studied by Sivinski et al. (1998, pers. comm.). This chapter examines similar dynamics on a larger scale, by comparing parasitoid abundance among trees within LaBelle. Additionally, I examined the temporal dynamics of the population as a whole, both within and among years.


72





73


Materials and Methods


Fruit Sampling in 1996


Loquat (Eriobotryajaponica (Thunb.)) and Surinam cherry (Eugenia uniflora L.) fruits were sampled at LaBelle and the adjacent town of Ft. DeNaud every two weeks from Week 4 (late January) to Week 22 (early June) of 1996. Most loquats were sampled from Week 4 to Week 14, but some were available until Week 18. Most Surinam cherries were sampled from Week 14 to Week 22, with one sample each collected in Weeks 4 and 6. Every tree within the towns which was found to have at least ten fruits was sampled. Each sample included fruits from a single tree. A total of 256 samples was collected, ranging from 3 in Week 4 to 51 in Week 16. Numbers of fruits per sample ranged between 17-106 for loquat, and between 18-151 for Surinam cherry. These numbers represented either all fruits present on the tree, or the maximal number that could be put in a single layer on the screen within the bucket (see Chapter 3 for details of fruit handling after collection). Fruits were sampled randomly from different parts of the tree. Each fruit sample was weighed following collection. Upon intensive collection of Surinam cherry fruits, it became apparent that they were infested by large numbers of fungal spores. Therefore, beginning in Week 18, fruits were washed in a 0.03% solution of sodium benzoate.

Abundance of the three parasitoid species, D. areolatus, D. longicaudata and Utetes anastrephae, was compared between LaBelle and Ft. DeNaud. Due to significant differences between the two towns (see Table 4-1), all subsequent analyses included only





74


data from LaBelle. Additionally, because U. anastrephae was relatively uncommon at LaBelle, in was not considered in these analyses. Analysis of Distribution Among Trees


Spatial distributions of Caribbean fruit flies, Anastrepha suspensa (Loew), and their parasitoids among trees were examined using data collected in LaBelle during 1996. Distribution was visualized using the Surfer software program (Golden Software) and the kriging method. Kriging uses sampled data to produce a grid of estimated values quantifying the entire distribution of the parameter of interest. Ultimately, kriged data are used to create isolines of equal parameter density visualized as a 2-dimensional contour map. For a detailed description of spatial analysis and its use in entomology see Brenner et al. (1998). Longitude and latitude coordinates for each host tree were obtained using the Microsoft Automap Street Plus software program, and adjusted slightly to fit a TIGER/Line base map (U.S. Census Bureau).

Figure 4-1 illustrates the quadrants which are included in this analysis. The town of LaBelle, situated to the south of the Caloosahatchee River, can be divided into four quadrants by State Road 80 transecting from east to west, and by State Road 29 running from north to south. A fifth section is in the town of North LaBelle, to the north of the river. Several samples collected in other quadrants were not included in this analysis.

Because of the low occurrence of parasitoids in loquat, only Surinam cherry samples were considered in the spatial analysis. Separate maps were produced for each sampling period, from Week 14 to Week 22. Relationships within each sampling period between fly abundance and that of each parasitoid species, and between parasitoid species, were examined using regression or correlation analysis.










low uda


NodthIae








s .. . ....s. t


Figure 4-1. Map of LaBelle, Florida, showing quadrants sampled in study.


75


CaioosAtchee River









StNteRadvas S~ od8





76


Temperature Measurements

Parasitoid abundance may be influenced by local variability in temperatures. In order to assess the occurrence of spatial variability in winter temperatures, five Optic StowAway data loggers (Onset Computer Corp.) were placed from December 1996 through March 1997, three at LaBelle and two at Ft. DeNaud. Loggers were placed one meter above the ground on the northern side of major limbs of large Surinam cherry trees, from which a significant number of parasitoids had been recovered during the previous spring. Loggers were set to record the temperature every 24 minutes. Mean, mean minimum and mean maximum temperatures were calculated for each month. These variables were subjected to Friedman's two-way analysis for block designs. This was achieved with the SAS software program by obtaining ranking among sites within days and then performing an analysis of variance on these ranks among sites (SAS Institute 1982). Means were subsequently compared with the Waller-Duncan k-ratio t test. In addition to the above mentioned variables, extreme minimum and maximum temperatures were noted.


Comparisons of Parasitoid Abundance Among Years


Various studies on parasitoids of A. suspensa have been conducted at LaBelle during recent years. In addition to data from the current study from 1995 (see Chapter 3) and 1996, Sivinski et al. (1996, 1998) sampled parasitoids for various purposes during each of the years 1991 through 1994. Thus, comparisons of parasitoid abundance among years could be made, and relationships with environmental factors examined.





77


In 1991 and 1992, samples were collected in a fashion similar to the present study; No more than one sample was collected from each tree in a single week. No manipulations (besides transformation) were performed on these data prior to analysis.

In 1994, multiple samples were collected from each tree in a single day, and often trees were sampled more than once a week. In order to make reliable comparisons with other years, these data were manipulated in the following manner. Samples collected from a single tree on a single day were combined by summing the numbers of parasitoids and flies emerging. Parasitism levels were calculated for each tree each day. Where trees were sampled more than once a week, mean parasitism per tree per week was calculated, and was considered to be a single sample for analysis.

In 1993, two separate studies were performed, one as in 1991-1992 (single sample per tree per week) and the other as in 1994 (multiple samples per tree). Where multiple samples were collected, data were manipulated as with the 1994 data. The resulting parasitism levels per tree per week were considered single samples, and given equal weight in the analysis as samples from the other study.

Note that, as mentioned above, only samples from LaBelle were considered. Thus, the 1995 data used in this chapter is a subset of the "LaBelle" data given in Chapter 3, which includes Ft. DeNaud.

Associations of parasitoid abundance with environmental factors were examined by linear regression analysis. Temperature and precipitation data were obtained from the Southeast Regional Climate Center, Columbia, South Carolina. Parasitism by each parasitoid species in loquat or Surinam cherry was related with the following variables: mean temperature, mean and extreme minimum temperatures, mean and extreme





78


maximum temperatures, and precipitation. Separate analyses were performed for conditions prevalent during each month preceding fruit collection. Thus, parasitism in loquat was related with temperature and precipitation variables for the months October through February, and parasitism in Surinam cherry was related with these variables for the months October through April.


Results


Comparison of Abundance Between LaBelle and Ft. DeNaud


The various parasitoid species varied greatly in their abundance between LaBelle and Ft. DeNaud during 1996 (Table 4-1). Parasitism by D. areolatus in loquat averaged 3% at LaBelle, but it was not collected at Ft. DeNaud. More D. longicaudata were recovered from loquat at Ft. DeNaud than at LaBelle, but the difference was not significant. U. anastrephae was uncommon in loquat at both towns. The differences between the towns were pronounced in Surinam cherry; D. areolatus was more common at LaBelle than at Ft. DeNaud, while both D. longicaudata and U. anastrephae were more abundant at Ft. DeNaud than at LaBelle. Note that these towns border each other, and the eastern most sample included in this study from Ft. DeNaud is only 2.2 km away from the western most sample from LaBelle. Because of these differences between the towns, only data from LaBelle were included in subsequent analyses.





79


Table 4-1. Comparisons of percent parasitism by various parasitoid species in 1996 between the towns of LaBelle and Ft. DeNaud. Fruit Parasitoid LaBelle Ft. DeNaud t p
species n Mean (SE) n Mean (SE)

Loquat D. areolatus 50 2.87(1.02) 29 0 3.58 0.001
D. longicaudata 50 0.09 (0.07) 29 0.61 (0.37) -1.43 0.16 U. anastrephae 50 0.02 (0.02) 29 0.03 (0.03) -0.30 0.77 Surinam D. areolatus 117 30.15 (3.02) 20 5.87(3.51) 5.12 0.001 cherry D. longicaudata 117 1.86 (0.49) 20 11.18 (3.57) -3.02 0.007
U. anastrephae 117 0.21 (0.09) 20 3.66 (1.52) -2.41 0.026


Distribution Among Trees


Densities of A. suspensa were generally low at the beginning of the Surinam cherry season. During Week 14, flies were concentrated at two focal points, one in the northeastern quadrant of LaBelle, and the second in North LaBelle, just north of the Caloosahatchee River (Figure 4-2). During Week 16, fly densities were again high at these locations. However, high infestations were observed also just to the south of the river, in the southern region of the northwestern quadrant, and at the southern fringes of town (Figure 4-3). Infestation levels were generally higher during Week 18, with highest numbers observed in North LaBelle, the northeastern quadrant just south of the river, and the northwestern quadrant (Figure 4-4). During Week 20, fly numbers were high at most locations, with highest infestations observed in the southwestern quadrant (Figure 4-5). A similar pattern was observed during Week 22, with focal points in the southwestern and northwestern quadrants (Figure 4-6).





80


D. areolatus was uncommon at the beginning of the Surinam cherry season. During Week 14, only small numbers were recovered from two host trees (Figure 4-7). Note that these were the same trees which had the highest infestations of A. suspensa (Figure 4-2). By Week 16, D. areolatus was recovered from 11 of 35 host trees. Four focal points were observed, in the northwestern, northeastern and southwestern quadrants (Figure 4-8). Parasitism levels increased dramatically by Week 18, with highest levels (over 50%) observed in the general vicinity of the focal points in the previous sampling period (Figure 4-9). Parasitism of over 50% was widespread during week 20, reaching over 80% in three areas: North LaBelle, the northeastern and extreme northwestern quadrants, and the southwestern quadrant (Figure 4-10). A similar pattern of generally high parasitism levels was observed during Week 22 (Figure 4-11).

D. longicaudata was not recovered from Surinam cherry at LaBelle until Week 16, when it was found in a single host tree just south of the river (Figure 4-12). During Week 18 it was recovered from the same tree, and one additional tree in the southwestern quadrant (Figure 4-13). Parasitism increased dramatically by Week 20, when it was recovered from 11 of 28 hosts. Four focal points with over 10% parasitism were apparent, two each in the northwestern and southwestern quadrants (Figure 4-14). During Week 22, highest parasitism levels were observed in the southwestern quadrant, with a second area of parasitism apparent on both sides of the Caloosahatchee River (Figure 4-15).

Parasitism by D. areolatus was significantly related with A. suspensa infestation levels during Week 14 (R2=0.68, F=10.86, p=0.02) and Week 16 (R2=0.45, F=23.67, p<0.0001), but not Week 18 (R2=0.1l, F=2.73, p=0.ll), Week 20 (R2=0.07, F=1.95, p=0.17), or Week 22 (R2=0.02, F=0.19, p=0.67). Parasitism by D. longicaudata was not





81


significantly related with A. suspensa infestation levels during any week (Week 20, R=0.03, F=0.91, p=0.35; Week 22, R2= 0.10, F=1.09, p=0.32).

The ratio between parasitism by D. longicaudata and that by D. areolatus mirrors the parasitism by D. longicaudata alone during Weeks 20 and 22 (Figures 4-16 and 4-17, compare with Figures 4-14 and 4-15). This is probably due to the relatively even distribution of D. areolatus during these weeks (Figures 4-10 and 4-11). Parasitism by the two species was significantly correlated during Week 18 (R=0.44, p<0.03). With all samples considered, there were no significant relationships between parasitism levels of D. areolatus and D. longicaudata during any other week (Week 16, R=0.10, p=0.60; Week 20, R=0.03, p=0.87; Week 22, R=0.28, p=0.35). However, when considering only samples from which D. longicaudata was recovered, there was a significant negative relationship during Week 20 (Figure 4-18).


Temperature Measurements


Considering the apparent negative relationship between cold winter temperatures and presence of this species (Chapter 3), it was hypothesized that the abundance of D. longicaudata at Ft. DeNaud relative to LaBelle may be the result of warmer winter temperatures. It was furthermore hypothesized that the river may have a moderating effect on temperatures. Temperature variables obtained at five sites at LaBelle and Ft. DeNaud are detailed in Table 4-2. The three coldest locations in all months, in terms of extreme minimum temperature, were the two locations at Ft. DeNaud and the site at LaBelle closest to the river. In terms of mean minimum temperatures, both Ft. DeNaud sites were significantly colder than two of the three LaBelle sites (Table 4-3). Contrary to expectations, the LaBelle site farthest from the river was the warmest location in terms of





82





26.775-~





26.77026.765-7




26.76026.755--81.445 -81.440 -81.435 -81.430 -81.425
Longitude (degrees)


0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Pupae / gram fruit Figure 4-2. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 14th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 7 of I 1 hosts.





83


26.775 26.770


U,

0, G)
*0

~0

(U
-J


-81.445 -81.440


-81.435


-81.430


-81.425


Longitude (degrees)



0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Pupae / ram fruit





Figure 4-3. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 16th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 31 of 35 hosts.


r


26.765 26.760


26.755





84


26.775





26.770


-81.445 -81.440


/--


-81.430


-81.425


Longitude (degrees)


0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Pupae / gram fruit




Figure 4-4. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 18th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 24 of 25 hosts.


-81.435


(u

-j


26.765 26.760


26.755










26.775 26.770


-81.445 -81.440
Lon


0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Pupae / gram fruit




Figure 4-5. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 20th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 28 of 28 hosts.


85


-81.435
gitude (degrees)


F-


-o


26.765 26.760


26.755


-81.430


-81.425


e






86


26.775





26.770


V


-81.445 -81.440


Longitude (degrees)


0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Pupae / gram fruit




Figure 4-6. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 22nd week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 14 of 14 hosts.


-81.435 -81.430 -81.425


26.765


Cl) 0) 0)
-0



__j


26.760 26.755





87


26.77526.770U)

26.76526.76026.755-


-81.445


-81.440


-81.435


-81.430


-81.425


Longitude (degrees)


0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Parasitism level





Figure 4-7. Spatial distribution of parasitism by Doryclobracon areolatus in Surinam cherry fruits at LaBelle during the 14th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 2 of 11 hosts. Parasitism level is the ratio between the number of D. areolatus emerging and the number of all parasitoids and flies emerging.


_


_ __L_ -





88


U)




-o


26.775 26.770 26.76526.76026.755-


-81.445


-81 .440


-81.435


-81.430


~


Longitude (degrees)



0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Parasitism level


Figure 4-8. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 16th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 11 of 35 hosts. Parasitism level is the ratio between the number of D. areolatus emerging and the number of all parasitoids and flies emerging.


-81.425





89





26.775CD26.7650)


-1

26.76026.755-81.445 -81.440 -81.435 -81.430 -81.425
Longitude (degrees)


0. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Parasitism level




Figure 4-9. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 18th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 15 of 25 hosts. Parasitism level is the ratio between the number of D. areolatus emerging and the number of all parasitoids and flies emerging.





90





26.775




0,0
26.77026.765





26.76026.755





-81.445 -81.440 -81.435 -81.430 -81.425
Longitude (degrees)


0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Parasitism level Figure 4- 10. Spatial distribution of parasitism by Doryclobracon areolatus in Surinam cherry fruits at LaBelle during the 20th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 27 of 28 hosts. Parasitism level is the ratio between the number of D. areo/atus emerging and the number of all parasitoids and flies emerging.





91





26.77 5





26.775





S26.765



ca
26.76026.755
0.&











-81.445 -81.440 -81.435 -81.430 -81.425
Longitude (degrees)



0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Parasitism level



Figure 4-11. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 22nd week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 13 of 14 hosts. Parasitism level is the ratio between the number of D. areolatus emerging and the number of all parasitoids and flies emerging.





92


26.7751 26.77026.765-26.76026.755-


-81.445


-81.440


-81.435 -81.430 -81.425


Longitude (degrees)


0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Parasitism level


Figure 4-12. Spatial distribution of parasitism by Diachasmimorpha longicaudata in Surinam cherry fruits at LaBelle during the 16th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from I of 35 hosts. Parasitism level is the ratio between the number of D. longicaudala emerging and the number of all parasitoids and flies emerging.


0)
CD 4)




Full Text

PAGE 1

BIOGEOGRAPHY OF BRACONID PARASITOIDS OF THE CARTOBEAN FRUIT FLY, ANASTREPHA SUSPENSA (LOEW) (DIPTERA: TEPHRITIDAE), IN FLORIDA AVRAHAM EITAM 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 1998

PAGE 2

1 ACKNOWLEDGEMENTS i i Fruits were collected and maintained by the Florida Department of Agriculture ! and Consumer Services, Division of Plant Industry (DPI), Bureau of Plant and Apiary Inspection. Many thanks to bureau chief Richard Clark, regional supervisors Terry Kipp ^ and Debra Chalot, and the numerous bureau personnel who assisted in this project. Additional thanks are due to Calie Jenkins and Joyce Willis for their assistance to field studies, and to Sam Simpson for the use of his laboratory to hold fiuit samples. Additional collections were performed by the U.S. Department of Agriculture, Animal and Plant Heahh Inspection Service (USDA-APHIS). Thanks to John Thomas, Earl Wiley and Ralph Cooley. Parasitoids and host flies were supplied by the DPI Caribbean Fruit Fly Mass Rearing Facility. Thanks to Don Harris, chief of the Bureau of Methods Development and Biological Control, and to Ed Bums, Suzanne Fraser, Mary Jo Hayes and other facility personnel for their assistance. Tracy Austin, North Florida Research and Education Center, Quincy, Florida, supplied me with Caribbean fhiit fly trapping data. Richard Brenner, David Milne, Jon Allen and Carlyle Brewster assisted in the preparation of contour maps. Yoav Gazit, Ali Harari and Kevi Vulinec cared for the laboratory cultures while I was away on field trips. ii

PAGE 3

Thanks to my supervisory committee members, Pat Greany, Howard Frank, Jonathan Crane and Richard Baranowski, for there useful suggestions. Special thanks to the committee chair, John Sivinski, for his assistance, support and patience. Many thanks to Tim Holler, USDA-APHIS, for his assistance throughout the course of this dissertation. Among other things, Tim helped organize fruit collections, participated in the host plant surveys, and assisted in maintenance of laboratory cultures. Many of my insights regarded parasitoid distribution were bom from hours of discussion with Tim during our travels together. Thanks to my friends and colleagues at the Department of Entomology and Nematology and at the U.S. Department of Agriculture, Agricultural Research Service, who helped me maintain some degree of sanity. Finally, thanks to my parents, whose many years of love, support and understanding have enabled me to achieve my goals. iii

PAGE 4

TABLE OF CONTENTS Elge ACKNOWLEDGMENTS ii ABSTRACT vii CHAPTER 1 GENERAL INTRODUCTION: THE CARIBBEAN FRUIT FLY AND ITS PARASITOIDS IN FLORIDA 1 CHAPTER 2 LITERATURE REVIEW: DISTRIBUTION, TEMPERATURE TOLERANCE AND DIAPAUSE OF PESTIFEROUS TEPHRITID FRUIT FLIES AND THEIR PARASITOIDS IN TROPICAL AND SUBTROPICAL REGIONS 5 Distribution and Population Dynamics 5 Hawaii 5 Anastrepha and Ceratitis in Tropical and Subtropical America 13 Bactrocera oleae in Southern Europe 18 Bactrocera tryoni in Australia 18 Conclusion 19 Effects of Temperature and Occurrence of Diapause 20 CHAPTER 3 LARGE-SCALE DISTRIBUTION PATTERNS OF CARIBBEAN FRUIT FLY PARASITOIDS IN FLORIDA 23 Materials and Methods 24 Fruit Sampling 24 Abiotic Environmental Data 29 Host Fly and Host Plant Data 31 Statistical Analysis 33 Results 35 Distribution and Abundance of Parasitoids 35 Host Plant Density 45 iv

PAGE 5

Relationships With Environmental Factors 52 Discussion 55 CHAPTER 4 LOCAL TEMPORAL AND SPATIAL DISTRTOUTION PATTERNS 0• DIACHASMIMORPHA LONGICAUDATA AND DORYCTOBRACON AREOLATUS IN AN AREA OF CO-OCCURRENCE 72 Materials and Methods 73 Fruit Sampling in 1996 73 Analysis of Distribution Among Trees 74 Temperature Measurements 76 Comparisons of Parasitoid Abundance Among Years 76 Results 78 Comparison of Abundance Between LaBelle and Ft. DeNaud 78 Distribution Among Trees 79 Temperature Measurements 81 Comparisons Among Years 101 Discussion 112 CHAPTER 5 EFFECTS OF TEMPERATURE ON IMMATURE DEVELOPMENT, ADULT LONGEVITY AND OVIPOSITIONAL ACTIVITY IN DIACHASMIMORPHA LONGICAUADATA 1 16 Materials and Methods 116 Immature Development 116 Aduh Longevity and Ovipositional Activity 1 18 Results 119 Immature Development 1 19 Aduh Longevity and Ovipositional Activity 120 Discussion 131 CHAPTER 6 LABORATORY REARING OF DORYCTOBRACON AREOLATUS 134 Materials and Methods 134 Insects 134 Cage Setup 135 Exposure to Host Larvae 135 Immature Stages and Adult Emergence 136 Life History Traits 137 Results and Discussion 138 v

PAGE 6

CHAPTER 7 BEHAVIORAL RESPONSE TO HOST CHEMICAL CUES BY FEMALES OF DORYCTOBRACON AREOLATUS 143 Materials and Methods 143 Insects 143 Experimental Design 144 Response Variables and Statistical Analysis 145 Results 146 Discussion 147 CHAPTER 8 SUMMARY AND CONCLUSIONS 152 APPENDIX NUMBERS OF SAMPLES COLLECTED AND INSECTS EMERGING FOR VARIOUS SITES BY MONTH, YEAR AND FRUIT TYPE 155 REFERENCES CITED 171 BIOGRAPHICAL SKETCH 183 vi

PAGE 7

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 BIOGEOGRAPHY OF BRACONID PARASITOIDS OF THE CARIBBEAN FRUIT FLY, ANASTREPHA SUSPENSA (LOEW) (DIPTERA: TEPHRITIDAE), IN FLORIDA By Avraham Eitam May 1998 Chairman: Dr. John Sivinski Major Department: Entomology and Nematology Host fruits of the Caribbean fruit fly, including loquat, Surinam cherry, Cattley guava and common guava, were collected throughout central and southern Florida. Three species of braconid parasitoids were recovered. Diachasmimorpha longicaudata (Ashmead) was limited mostly to southern Florida, reaching higher latitudes along both coasts. Doryctobracon areolatus (Szepligeti) was common at most interior locations, but absent or rare along the coasts. Distribution of these two species overlapped only within a limited region, and only at LaBelle (Hendry Co.) did they commonly co-occur. Utetes anastrephae (Viereck) was widespread, but its abundance was inversely related with that of D. areolatus. Absence of D. longicaudata was related with low temperatures, but was best explained by high variability of temperatures. Two hypotheses are proposed to explain the relationship between temperature and D. longicaudata distribution: (1) Low temperatures have a direct negative effect; (2) Variable or low temperatures adversely vii

PAGE 8

affect host availability, which in turn has a negative effect on D. longicaudata. Evidence supporting each hypothesis is discussed. Parasitism levels by D. longicaudata in loquat and common guava fhiits were significantly related with the minimum and mean numbers, respectively, of Caribbean fruit flies captured in McPhail traps. Similarly, parasitism by U. anastrephae in loquat and Surinam cherry fruits was related with minimum fly numbers. Parasitism levels of all species combined in loquat and Surinam cherry fruits was significantly related with densities of common guava trees. Parasitism by U. anastrephae in Cattley guava fruits was related with densities of Surinam cherry plants. The apparent disappearance of D. areolatus from the southern Atlantic coast, where it was originally released, may be partially due to interspecific competition. Mechanisms proposed that may give D. longicaudata a competitive advantage include better ability to locate larvae within fruits, a longer ovipositor allowing greater access to hosts, higher fecundity, and an advantage in competition among larvae. At LaBelle, parasitism by D. longicaudata in spring fruiting loquat and Surinam cherry was positively related with the preceding winter temperatures. A similar relationship was found for D. areolatus, but only in loquat fruits. A negative relationship between parasitism by D. areolatus and D. longicaudata was observed at the peak of the Surinam cherry fruiting season, suggesting that significant competition may occur. viii

PAGE 9

CHAPTER 1 GENERAL INTRODUCTION: THE CARIBBEAN FRUIT FLY AND ITS PARASITOIDS IN FLORIDA The Caribbean fruit fly, Anastrepha suspensa (Loew), became established in Florida in 1965, quickly spreading throughout southern and central Florida (Weems 1966). In Indian River County, A. suspensa occurrence was linked primarily with the availability of various host fruits (Nguyen et al. 1992). Flies were collected mainly from loquat (Eriobotrya japonica (Thunb.)) during December-April, Surinam cherry {Eugenia uniflora L.) dining MayJune, and Cattley guava (Psidium cattleianum Sabine) during July-August, with greatest numbers reported from the latter two fruits. A population increase during one of the survey years in November-December was related to a second crop of Surinam cherry. In Dade County, weekly trap catches appeared to mirror temperature fluctuations (Hennessey 1994). Correlations of fly catches with rain and ten:q)erature together were significant for most years. Hennessey (1994) concludes that abiotic environmental factors and host availability interact to affect trapping frequency. In an effort to control A. suspensa, several species of parasitoids were introduced to Florida (Baranowski et aL 1993). The first to be released was Doryctobracon areolatus {Parachasma cereus) (Szepligeti) (Hymenoptera: Braconidae: Opiine) (Baranowski and Swanson 1970). This is a widespread larval-pupal species, ranging from Mexico to Argentina (Wharton and Marsh 1978). An origmal stock of 7 males and 17 females from Trinidad was reared through 6-7 generations, and 45 males and 26 females were released 1

PAGE 10

2 at Homestead in 1969 (Baranowski et al. 1993). Although it was recovered in large numbers the following simimer, populations at this site have since declined (Baranowski et al. 1993). Small numbers persisted in the area at least until the occurrence of Hurricane Andrew in 1992 (Sivinski 1991, pers. comm.). Studies indicate that D. areolatus abimdance varies among locations in Florida. Sivinski et al. (1996) report it to be the dominant parasitoid in areas west of Lake Okeechobee. However, it was absent from their samples in southeastern Florida. Holler (unpublished data) felled to recover this species in a survey preceding augmentative releases of Diachasmimorpha longicatidata (Ashmead) along the central Atlantic coast of the state. However, subsequent intensive fruit sampling during 1993-1994 produced 16 D. areolatus from 7 trees (Denise Marshall, pers. comm.). D. longicaudata, another larval-pupal opiine braconid, was introduced in 1972. This Indo-Philippine species was originally recovered from Bactrocera ^p. (Clausen 1978). It has been utilized in the biological control of a wide range of tephritid hosts in various regions of the world (see Chapter 2). In contrast to the limited release of D. areolatus, D. longicaudata was released in large numbers in 21 counties throughout central and southern Florida (Baranowski et aL 1993). Based on reduced fly catches in subsequent years, it appeared to have had a significant impact upon host fly populations (Baranowski et aL 1993). Two other exotic larval-pupal parasitoids, Aceratomuromyia indica Silvestri (Hymenoptera: Eulophidae) and Trybiographa daci Weld (Hymenoptera: Eucoilidae), also were considered established (Baranowski et al. 1993). In additbn two larval-pupal braconid parasitoids, Utetes anastrephae (Viereck) and Doryctobracon anastrephilum

PAGE 11

(Marsh), were recovered in small numbers prior to the parasitoid releases. These were considered to have originally existed on Anastrepha interrupta Stone in the Florida Keys (Baranowski et al. 1993). Like D. areolatus, U. anastrephae is a wide ranging species, distributed south to Argentina (Wharton and Marsh 1978). Sivinski et al. (1998) investigated the temporal dynamics of D. longjcaudata and D. areolatus populations at LaBelle. D. longjcaudata became more abundant, actually and relative to D. areolatus, as the season progressed, in all fruits except calamondin {Citrus mitis Blanco). Tenqjerature best explained the fluctuations in relative abundance. However, with the exception of the autumn-winter decline of D. longicaudata in calamondin, results could also be explained by "counter-balanced con^tition" (cf Zwolfer 1971), where D. areolatus is superior to D. longicaudata in finding host patches, but inferior at exploiting hosts. Augmentative releases of adult D. longicaudata apparently substantially suppressed A. suspensa populations at two locations in Florida (Sivinski et al. 1996). Similar results have been reported with releases of Diachasmimorpha tryoni (Cameron) for the suppression of the Mediterranean fruit fly, Ceratitis capitata (Wiedemann) in Hawaii and Guatemala (Wong et al. 1991, Sivinski et al. submitted). Inimdative releases ofD. longicaudata were being employed for A. suspensa control for several years in the central Atlantic coast region of Florida (Bvims et al. 1996). The objectives of this study were to detennine the current distribution patterns and relative abundance of A. suspensa parasitoids in Florida, and identify fectors affecting this distribution. These determinations could assist in the ongoing biological

PAGE 12

control effort, by suggesting which parasitoid species should be employed in augmentative releases in various regions of the state. The following chapter reviews literature on the distribution of tropical and subtropical tephritid fixiit flies and their parasitoids in other regions of the world. Chapter 3 describes the geogr^hic distribution of A. suspensa parasitoids in Florida, and the environmental fectors associated with this distributioa Chapter 4 investigates tenqwral and spatial dynamics in an area of co-occurrence of parasitoid species within Florida. Subsequent chapters investigate biological attributes >\iuch may influence parasitoid distribution and abundance. Chapter 5 examines effects of tenq)erature on D. longicaudata adults and immature stages in the laboratory. Chapter 6 describes life history traits of D. areolatus in the laboratory. Chapter 7 investigates the host location behavior of D. areolatus. Generic names of opiine braconids in this text are according to the recent revision by Wharton (1997).

PAGE 13

CHAPTER 2 LITERATURE REVIEW: DISTRIBUTION, TEMPERATURE TOLERANCE AND DIAPAUSE OF PESTIFEROUS TEPHRITID FRUIT FLIES AND TfflER PARASITOIDS IN TROPICAL AND SUBTROPICAL REGIONS Distribution and Population Dynamics A great deal of literature has been published concerning distribution of tephritid fruit flies and their parasitoids in various regions of the world. The two regions most extensively studied in this regard have been Hawaii and tropical America. The objective of this review is to identify fectors which may be important in explaining fruit fly and/or parasitoid distribution patterns. Hawaii Four species of adventive frugivorous tephritids occur in Hawaii. They are melon fly, Bactrocera cucurbitae (Coquillett), arrived in 1895; Mediterranean fruit fly, Ceratitis capitata (Wiedemann), in 1910; oriental fruit fly, B. dorsalis (Hendel), in 1945; and Malaysian fruit fly, B. latifrons (Hendel), in 1983 (Vargas et al. 1989). Shortly after its arrival, C. capitata became a serious economic pest on various fruits throughout Hawaii (Back and Pemberton 1918). Subsequent to the arrival of the B. dorsalis, C. capitata became scarce at lower elevations but remained abundant in upland areas (Bess 1953). It was speculated that C. capitata had been conq)etitively displaced by B. dorsalis, which was better adapted to the warmer climate of the lowlands (Haramoto 5

PAGE 14

6 and Bess 1970). In guava fruits, B. dorsalis larvae suppressed the development of larvae of the C. capitata by an unspecified mechanism (Keiser et al. 1974). Since guava is a major host at low altitudes, this could contribute to the displacement of C. capitata in these areas. However, C. capitata has not been totally displaced at low elevations. Bess (1953) and Keiser et al. (1974) noted that it remained common in preferred hosts, such as coffee, regardless of elevation. During the winter months of 1966-1968 (when total infestation was low) it outnumbered B. dorsalis in guava (Haramoto and Bess 1970). Vargas et al. (1983a) studied C. capitata distribution on Kauai. Fly abundance was inversely related with elevation, and also with rainfall. Large numbers emerged from peach, loquat, sandalwood, and coffee. Highest populations were in areas containing scattered strands on feral coffee. In newly planted coffee fields in lowland Kauai, C. capitata dominated B. dorsalis by the end of each of three seasons (Vargas et aL 1995). The authors suggest that infestation of an earlier stage of fruit ripeness and fester development of C. capitata larvae in coffee reduce competitive interactions with B. dorsalis larvae. Furthermore, they suggest that absence of large overstory trees and a scarcity of alternate hosts may limit B. dorsalis abimdance in monoculture coffee fields. In studies on Oahu, C. capitata was found to occur in larger numbers than B. dorsalis in feral coffee, but was less common on other hosts (Harris and Lee 1986, 1987). The abundance of coffee berries and distribution of other fruits apparently influenced fluctuations in C. capitata populations. Ramfall had an indirect effect on C. capitata dynamics by inducing coffee fruiting (Harris and Lee 1986). In the urban areas of Oahu, there was apparently a more direct effect of ramfall. Medflies were more common m dry.

PAGE 15

7 leeward areas, although the same host fruits were grown there as in wet, windward areas (Harris and Lee 1987). Population cycles of B. dorsalis have been associated with fruiting of common and strawberry guava (Newell and Haramoto 1968, Vargas et al. 1983b). Although abundance is negatively related with elevation on Kauai, it appears that the factor limiting distribution at high elevations is the relative scarcity of hosts (Vargas et al. 1983b). Similarly, high numbers of B. dorsalis in wet windward areas vs. dry leeward areas, and outside vs. inside production areas, corresponds with concentrations of wild guava (Vargas etal. 1989, 1990). The major hosts of B. cucurbitae in Hawaii include tomato and various species of wild and cultivated cucurbits (Harris et al. 1986). Fly abxmdance on Kauai is negatively related with elevation and rainfall. These relationships may be e>q)lained by host plant distribution: hosts are not found above 300 m, and grow better in drier areas (Harris et al. 1986). On Molokai, fly distribution is strongly related to that of the feral host bittermelon (Harris and Lee 1989). Similarly, abundance of B. cucurbitae in dry leeward areas vs. wet windward areas on Kauai is related to the distribution of bittermelon and spiny cucumber (Vargas et al. 1989). More B. cucurbitae were captured inside production areas, again related to abundance of host plants (Vargas et al. 1989, 1990). B. latifrons develops on a variety of solanaceous and cuciurbitaceous plants (Liquido et al. 1994). Although B. cucurbitae is the primary fruit fly on most species in these families, B. latifrons appears to outcompete other fruit fly species on several host plants that inhabit disturbed, abandoned fields and less managed ranch lands (Liquido et

PAGE 16

8 al. 1994). Populations are apparently affected by both temperature and rainfall Liquido et al. (1994) suggest that high rainfall excludes this fly from the windward side of Hawaii. Following the arrival of C. capitata, a project was undertaken to introduce fruit fly parasitoids to Hawaii. F. Silvestri (during 1912-1913) and D.T. Fullaway and J.C. Bridwell (in 1914) released several species, 5 of which became established: the opiine braconids Diachasmimorpha tryoni (Cameron), Opius humilus Silvestri, and Diachasmimorpha (Biosteres) fullawayi (Silvestri), the eulophid Tetrastichus giffardianus Silvestri, and the chacidid pupal parasitoid Dirhimis giffardii Silvestri (Clausen et al. 1965, Gilstrap and Hart, 1987). All species were from Africa, with the exception of D. tryoni from Australia. Initially, O. humilus became the dominant species, reaching maximum levels of parasitism in 1915. D. tryoni was dominant from 1916 onward, with maximum parasitism recorded in 1918. Total parasitism from 1914 to 1933 ranged between 24.9-56.4% (Willard and Mason 1937). O. humilus disappeared from Oahu in the late 1930s, but Clausen et al. (1965) reported it as abundant in the Kona section of the island of Hawaii, equaling or exceeding D. tryoni in coffee. Interestingly, O. humilus is not reported in more recent literature. The displacement of O. humilus by D. tryoni may be due to a competitive advantage by the strongly mandibulate larvae of the latter species (Pemberton and Willard 1918). After its establishment, T. giffardianus parasitized up to 25.3% of C. capitata larvae, averaging 6.3% between 1914-1933 (Willard and Mason 1937). Subsequent to Clausen et al. (1965), it had not been reported until Ramadan and Wong (1990) found it to be abundant in the Kula area of Maui. Purcell et al. (1994) found that while this

PAGE 17

species is absent from gnavas collected on the tree, it is common in fiiiit remaining on the ground for 4-9 days. Piircell (submitted) reports that another eulophid, Aceraneuromyia indica Silvestri, is established on all major islands, but appears to be less abundant than T. gijfardianus. Until the early 1950s, D. fullawayi was readily recovered from cofifee and peach (Bess et aL 1961). Haramoto and Bess (1970) found this species only in cofifee plantations in Kona, Hawaii. It has not been reported since. Between 1947-1952, many species of parasitoids were introduced to Hawaii for the control of B. dorsalis (Clausen et al. 1965). Of these only four species of opiine braconids became permanently established: Fopius (Biosteres) arisanus (Sonan), Diachasmimorpha longicaudata (Ashmead), Fopitis (Biosteres) vandenboschi (Fullaway), and Psyttalia incisi (Silvestri). There was an interesting succession of parasitoids between 1948-1950. Initially, D. longicaudata was the dominant species. During the late summer and fell of 1949, F. vandenboschi increased in abimdance, and by the end of the year had become fer more nimierous than D. longicaudata (Bess et al. 1950). During 1950, F. arisanus increased dramatically, and on Oahu constituted 99.4% of the total parasitism in December (van den Bosch et al. 1951). The total parasitism also increased, to approximately 80% (van den Bosch et al. 1951). F. arisanus has remained the dominant parasitoid ever since. Several factors may explain this successive displacement. F. arisanus attacks the egg of its host (van den Bosch and Haramoto 1951), F. vandenboschi first-instar larvae, and D. longicaudata second and third-instar larvae. While all eggs are accessible to F. arisanus, some larvae may escape parasitism by burrowing into the fruit pulp, especially

PAGE 18

10 in large inxrt (Sivinski 1991). This phenomenon would increase with older larvae, thus putting D. longicaudata at a disadvantage. Additionally, F. arisanus larvae inhibit the development of F. vandenboschi and D. longicaudata^ and those of F. vandenboschi inhibit the development of D. longicaudata (van den Bosch and Haramoto 1953). This inhibition is apparently by means of physiological suppression and not physical injury (van den Bosch and Haramoto 1953). The early predominance of D. longicaudata was apparently enhanced by the disproportionate release of large nimibers of individuals of this species (van den Bosch et al. 1 95 1 ). Note that Palacio et al. (1991) found no evidence of physiological suppression in a study on competition among F. arisanus, D. longicaudata and Fopius (Biosteres) persulcatus Silvestri. Instead, they foimd that D. longicaudata was a superior conpetitor to both F. arisanus and F. persulcatus, with F. persulcatus being superior to F. arisanus, indicating physical competition among first-instar larvae. They further reported that D. longicaudata did not discriminate between parasitized and unparasitized hosts, while F. persulcatus avoided superparasitism. However, Lawrence et al. (1978) demonstrated that D. longicaudata does avoid superparasitism when provided with large niraibers of hosts. F. arisanus, D. longicaudata, and F. vandenboschi develop not only on B. dorsalis, but also on C. capitata. The reduced abundance of C. capitata in the early 1950s was partially attributed to the effects of these parasitoids, particularly F. arisanus (Bess 1953). P. incisi does not develop on C. capitata (Stark et al. 1994). D. tryoni does not develop on B. dorsalis, except in cases of multq)arasitism involving D. longicaudata (Ramadan et al. 1994a). However, it develops on two species

PAGE 19

11 of gall-forming tephritids, the eupatorium gall fly, Procecidochares utilis Stone, and the lantana gall fly, Eutreta xanthochaeta Aldrich (Haramoto and Bess 1970). F. arisanus is the dominant parasitoid of both B. dorsalis and C. capitata in Hawaii. However, other species may constitute a large part of the total parasitism imder certain circumstances. Wong et al. (1984) and Wong and Ramadan (1987) studied the parasitoid feuna on both species of fiiiit flies in the Kula area of Maui. In these studies, fly pupae were not separated, so the various parasitoid species could not be attributed to a specific fly species. D. longicaudata and D. tryoni were quite common on loquats and peaches, occasionally surpassing F. arisanus in abundance. For example, D. longicaudata accounted for 34.9% of the total parasitism in 1979 from peaches, and 33.4% in 1984 from loquats, and D. tryoni accoxmted for 32.7% in 1980 from loquats. P. incisi and F. vandenboschi accoimted for 1.6% and 0.2 of the total parasitism, respectively. At a site at 1200 m elevation on Hawaii island, D. tryoni was the dominant species, and often the only one recovered (M. Purcell pers. comm.). Dominance at high altitudes may be related to an observation by Pemberton and Willard (1918) that mature D. tryoni larvae enter a winter diapause within host puparia. Several additional studies have reported relative abundance of B. dorsalis parasitoids. Vargas et aL (1990) reported that in ripe fiiiit in an agricultural area, D. longicaudata and P. incisi constituted 3.9-5.5% and 0.4-2.3% of the total parasitism, respectively. Stark et al. (1991) determined the abundance of parasitoids in commercial guava by canopy fogging. D. longicaudata, P. incisi, and F. vandenboschi accounted for 9-10%, 2-10%, and 0.25-1% of the total parasitism, respectively. Vargas et al. (1993)

PAGE 20

12 report that these three species were more common in orchards than in wild guava. They suggest that two possible factors contributing to this observation may be high tree densities and abundance of rotting fruit in commercial guava orchards. Vargas et al. (1993) point out that fruit type influenced parasitoid abimdance. F. vandenboschi represented 8.8% of the total parasitoids collected from passion fruit in 1988. Small fruits such as Surinam cherry and felse kamani produced many P. incisi, while D. longicaudata was often common in mango (32.2% in 1988). Purcell et al. (1994) sampled guava fruit from the tree and the ground. They found that D. longicaudata abundance increased as fruit aged on the groimd, and P. incisi was recovered only from fruit on the ground at least 4 days. This suggests that these parasitoids forage on fruit on the groimd, and sampling fruit only from the tree would underestimate their abundance. F. vandenboschi abundance (less than 3% of total parasitoids) was imaflfected by fruit ripeness. Vargas et al. (1995) studied abundance of Mediterranean and oriental fruit flies and their parasitoids in newly-planted coffee fields. Although parasitism of C. capitata by F. arisanus was apparently density-dependent, low parasitism (33.1-37.6%) was observed. Interestingly, F. arisanus parasitized a greater percentage of C. capitata, the more common host, than B. dorsalis. On Oahu, F. arisanus appeared to be inefficient in parasitizing hosts at low population densities (Harris and Lee 1987). The opiine braconid Psyttalia fletcheri (Silvestri) was introduced in 1915-1916 for control ofB. cucurbitae (Clausen et al. 1965). It is the only significant parasitoid of B. cucurbitae in Hawaii (Nishida 1955). This species was more common on the wild hosts Momordica balsamina L. and M. charantia L. than on cultivated hosts, with parasitism

PAGE 21

13 levels of up to 50% during favorable seasons (Nishida 1955). Parasitism was highest in winter and lowest in summer (Nishida 1955). Larvae in vines were more highly parasitized than larvae in fruits, possibly because in the latter they could escape parasitism by penetrating deeply into the pulp (Nishida 1955). This may also explain the low parasitism on large cultivated fruits. Parasitism in cultivated fruits was higher in very weedy fields, suggesting that P.fletcheri favors weedy situations (Nishida 1955). Harris and Lee (1989) suggest that the absence of P. fletcheri from Molokai may be due to imfavorable high winds on that island. Liquido et al. (1994) report very low (less than 1%) parasitism of B. latifrons by D. longicaudata and Tetrastichus sp, Anastrepha and Ceratitis in Tropical and Subtropical America The genus Anastrepha includes 184 described species ranging from the southern United States to northern Argentina (Aluja 1994). At least 54 species occur in Panama (Stone 1942) and 23 in Mexico (Aluja et al. 1987). Biological knowledge is basically restricted to seven economically important species: fraterculus, grandis, ludens, obliqua, serpentina, striata, and suspensa (Aluja 1994). Several studies of Anastrepha abimdance have been conducted in Chiapas, southern Mexico. Celedonio-Hurtado et al. (1995) trapped flies in orchards of various fruit species. Fruit fly species composition varied among orchards, with 1 or 2 predominant species representing 43-86% of all individuals. For example, in sapodilla, Achras zapota L., 86% of all flies trapped were A. serpentina (Wiedemann), while in chalum, Inga micheliana Harms, 66% were A. distincta Greene and 25% A. ludens (Loew). RainfeU could not explain population fluctuations, and the authors conclude that

PAGE 22

14 host fruit availability is the most important fector affecting adult populations. A. obliqm (Macquart) and A. ludens are the predominant species in mango, with A. ohliqua being more common at lower elevations and A. ludens at higher elevations (Aluja et al. 1987, 1990, 1996). In another study conducted in a coffee producing area of the same state, A. ludens was the most abimdant species with 60% of trapped flies, followed by A. distincta and A. fraterculus (Wiedemann) with 22 and 12%, respectively (Malo et al. 1987). The occurrence of the latter two species was related to the abimdance of Inga spp., the main host oiA. distincta, and coffee, a minor host oi A. fraterculus, in the study area. Studies in other countries showed similar tendencies. In Costa Rica, A. obliqua was associated with mango and other Anacardiaceae, A. striata Schiner with guava and other Myrtaceae, and A. serpentina with species of Sapotaceae (Jiron and Hedstrom 1988, 1991). Soto-Manitiu and Jiron (1989) found that the maximum abimdance of each species coincides with the fruiting season of their respective host plants. Most A. obliqua emerge just after first rains, coinciding with the mango fruiting seasoa In citrus orchards in Belize, the seasonal increase in numbers of A. ludens trapped was derived mainly from infestations in grapefruit (Houston 1981). In Brazil population dynamics of A. fraterculus has been related to host fruit availability (Malavasi and Morgante 1981). Nascimento et al. (1982) report that A. obliqua was predominant in citrus orchards, and A. fraterculus in locaUties with tropical hosts, especially guava. While the occurrence of flies in citrus was related to host availability, no such relationship was observed with tropical hosts. Trapping was related to mean and minimum temperature and relative humidity. Fehn (1982) studied the population dynamics of Anastrepha spp. in peach orchards at three locations in two

PAGE 23

15 seasons. He found relationships with various meteorological factors, including tenq>erature, relative humidity, rainfall and wind velocity, at some locations and seasons but not others. However, he suggests that availability of alternative hosts may be the principal factor affecting population dynamics. C. capitata invaded Costa Rica in 1955 and has since spread to all of Central America (references in Wharton et al. 1981), and into South America to Brazil (Aguiar and Menezes 1996). It comprised 5 and 19% of fruit flies collected in Costa Rica and Brazil, respectively (Jiron and Hedstrom 1988, Aguiar and Menezes 1996). Several species of parasitoids were released for the control of C. capitata in Central America (Gilstrap and Hart 1987). Of these, three species—Z). longicaudata, F. arisanus and the eulophid A. indica—yfere recovered by Wharton et al. (1981) in Costa Rica. In this study, D. longicaudata and F. arisanus were the dominant parasitoids of C. capitata, while A. indica and D. longicaudata were the most common on Anastrepha spp. Native parasitoids occurred in much smaller numbers. For example, parasitism of Anastrepha spp. by the opiines Doryctobracon areolatus (SzepUgeti) and Utetes anastrephae (Vier.) was only 0.2 and 0.05%, respectively. In a later study, Jiron and Mexzon (1989) report that D. areolatus was the most abundant and widespread species. In Guatemala, D. longicaudata was reported to be the most common parasitoid of C. capitata, while that of Anastrepha spp. was Doryctobracon crawfordi (Viereck), followed by D. areolatus and U. anastrephae (Eskafi 1990). The combined parasitism in this study was very low (<2%) for all fruits except Surinam cherry with 8% parasitism Various exotic fruit fly parasitoids were introduced into Mexico in the 1950s (JimenezJimenez 1956, 1958). Of these, D. longicaudata wad A. indica were estabUshed

PAGE 24

16 (Clausen 1978). Several systematic surveys of Anastrepha parasitoids were subsequently conducted, producing very different results concerning the relative abimdance of parasitoid species. In an area of mixed cultivation in the State of Chiapas in southern Mexico, Aluja et al. (1990) found that the most abundant parasitoid was D. longicaudata. However, in a native tropical community in the State of Veracruz, D. areolatus and U. anastrephae represented 59 and 17%, respectively, of the total parasitism, while D. longicaudata was not recovered at all (Hernandez-Ortiz et al. 1994). Lopez et al. (submitted) confirmed that D. areolatus is the most common species in Veracruz, representing 43% of all parasitoids recovered from various habitats. It also had the widest host breadth of all parasitoid species. In an earlier study in the State of Nuevo Le6n in northeastern Mexico, Gonzalez-Hernandez and Tejada (1979) reported that the most common parasitoid was D. crawfordi, followed by D. areolatus. Interestii^ly, in Veracruz D. crawfordi was common only in citrus (Lopez et al., submitted). Exotic parasitoids also were introduced into other locations in the Americas for the control of Anastrepha species. D. longicaudata was introduced to Trinidad in 1974 (Bennett et al. 1977). The following year it was the most common parasitoid recovered, surpassing the native D. areolatus. In Argentina, D. longicaudata and A. indica were reported as established (Ovruski and Fidalgo 1994). Several studies in Brazil report that D. areolatus was by far the most abundant parasitoid of Anastrepha species. Leonel et al. (1995) found that 70% of parasitoids emerging from san^les collected in 10 states were D. areolatus. The alysiine Asobara anastrephae (Muesebeck) was the second most common species with 19% of the total parasitism, and U. anastrephae the third most common with 10%. In the state of SSo

PAGE 25

17 Paulo alone, D. areolatus and U. anastrephae constituted 84 and 6% of all parasitoids, respectively (Leonel et al. 1995). In Itaguai, Rio de Janeiro, 89% of all parasitoids collected were D. areolatus, with U. anastrephae accounting for an additional 8% (Aguiar-Menezes and Menezes 1997). In Amazonas State, D. areolatus was found to be the dominant parasitoid in rural locations while Opius sp. nr. bellus dominated in urban areas (Canal D. et al. 1994, 1995). Finally, D. areolatus was the dominant parasitoid of Anastrepha zenildae Zucchi in the State of Rio Grande do Norte (Araujo et al. 1996). D. areolatus also was the most common parasitoid reported from Venezuela, accomting for 33% of the parasitism (Katiyar et al. 1995). U. anastrephae was the fourth most common species with 7%. Ovruski (1995) reports low levels of parasitism from Tucuman province, Argentina, with D. areolatus emerging from less than 2% of Anastrepha spp. puparia. Earlier studies by Nasca (1973) and Fem^dez de Ardoz and Nasca (1984) also reported D. areolatus (as Opius tucumanus or Doryctobracon tucumanus) from the same province. Sivinski et al. (1997) analyzed the distribution of parasitoids of Anastrepha spp. within tree canopies in Mexico. Several tendencies were reported. Parasitism by U. anastrephae was observed only in a narrow range of small host fruits. The efficiency (proportion of larvae attacked in a fruit) of D. longicaudata compared to that of other parasitoids increased with fruit size. Parasitism by D. areolatus decreased during fruiting periods of individual trees as the season changed from rainy to dry. Negative relationships in parasitism were observed between D. areolatus and U. anastrephae, while the introduced D. longicaudata and native D. crawfordi tended to overlap.

PAGE 26

18 Bactrocera oleae in Southern Europe The olive fly, Bactrocera oleae (Gmelin) is native to Africa and currently distributed throughout the Mediterranean basm and the Middle East (Clausen 1978). The braconid Psyttalia concolor (Silvestri) was introduced to Italy from North Africa in 1914, and again in 1917-1918, 1923 and 1934 (Clausen 1978). It was also established in France and Greece (Clausen 1978). Inundative releases have been performed at various locations and have proven to be quite successfiil in reducing fly populations (e.g., Monastero and Delanoue 1966, Kapatos et al. 1977). While it is established in southern Italy, atten^ts at establishment in more northern regions have feiled (e.g., Raspi and Loni 1994). Bactrocera tryoni in Australia The Queensland fruit fly, Bactrocera tryoni (Froggatt), is native to Australia. Toward the southern fringe of its permanent distribution, there was a significant correlation between summer rainfall and peak fly numbers (Bateman 1972). The efiect was thought to be mediated through a reduction in fecimdity and immigration, and high mortality among adults emerging through dry soil in dry years. Several species of exotic parasitoids were introduced to Australia (Snowball et aL 1962). Only F. arisanus persisted on the Australian mainland and D. longicaudata on Lord Howe Island, even thoi^ both species were initially established at both locations (Snowball and Lukins 1964). At most locations, native parasitoids were imcommon relative to F. arisanus. Snowball and Lukins (1964) suggest that low winter temperatures may Umit the distribution of F. arisanus in southern Australia. Snowball (1966)

PAGE 27

19 concludes that fectors other that host availability may account for lower and more variable parasitism at higher latitudes, again suggesting some tenq)erature efifect. Conclusion Factors affecting distribution of fruit flies and their parasitoids could be put into three main categories: abiotic fectors, host availability and con^etition. The most commonly mentioned abiotic factors are temperature and precipitatioiL Tenq)erature may be important at limiting distribution at high latitudes, as was suggested for F. arisanus in southern Australia. Winter temperatures may also limit parasitoid distribution in Florida, given that these parasitoids originate in tropical regions and may not be adapted to cold temperatures. The fector most commonly mentioned as affecting distribution and abundance of tropical firuit flies and their parasitoids is host fioiit availability. In addition to many studies from Hawaii and tropical America noted above. Tan and Serit (1994) reached similar conclusions regarding B. dorsalis in Malaysia. Even in cases in which population dynamics appear to follow temperature changes, several authors have suggested an indirect effect of ten^rature, through its influence in host availability. Interspecific interactions among parasitoids may be very complex. Examples of successive replacement of parasitoid species in Hawaii have been regarded as classic examples on competitive displacement. Mechanisms suggested have involved competition among larval stages within fruit, including either physical interaction among larvae or physiological suppression. Additionally, parasitoids attacking early stages replaced parasitoids attacking later stages. It was suggested that earlier stages are more

PAGE 28

20 exposed to parasitoids because they are situated closer to the fruit surface, thus giving the former species an advantage. This mechanism may be irrelevant in Florida, as the three parasitoid species present in the state, namely D. areolatus, D. longicatidata and U. anastrephae, apparently all attack late-instar larvae. Effects of Temperature and Occurrence of Diapause The distribution of tropical fruit flies in warm tenq)erate climates can be limited by temperature. Therefore, the determination of low temperature tolerance is important. Messenger and Flitters (1954) used environmental chambers to simulate the climates of locations in the continental United States. They determined that C. capitata, B. cucurbitae, and B. dorsalis could successfully reproduce in most of Florida and along the Gulf coast, and possibly southern California. Flitters and Messenger (1965) stated similar conclusions for A. ludens. Meats (1981) and O'Loughlin et al. (1984) determined that the Queensland fruit fly, Bactrocera tryoni (Froggatt), could establish permanent, lowdensity populations in southern Australia. LevyaVazquez (1988) used a degree-day model to estimate lower thresholds and thermal constants for A. ludens. The lower thresholds for the various immature stages ranged from 9.4-14.rC. Thomas (1997) found that pupal duration m the field closely fit this laboratory-based model. The puparial stage may be prolonged up to three months in the winter, but there was no evidence of a winter diapause. Larval development time was variable and did not agree well with the model. Christenson and Foote (1960) reviewed the occurrence of diapause m fruit flies. While diapause is typical of most North American Rhagoletis spp., most tropical and subtropical species are not known to undergo diapause.

PAGE 29

21 Prescott and Baranowski (1971) determined the tenperature tolerances for A. suspensa. Eggs foiled to hatch below 12°C or above 33°C. No emergence was observed at 10 and 12°C, although pupae were still viable at 12°C when the ejq)eriment was terminated. The calculated development threshold was 10°C and the optimal development temperature approximately 25 °C for all immature stages. Ashley et aL (1976) studied adult emergence of A. suspensa and D. longicaudata between 22-32°C. Both flies and parasitoids had high levels of mortality above 28°C. In contrast Darby and Kapp (1934) report that A. ludens has greater tolerance than its parasitoid D. crawfordi at both high and low tenqieratures. No emergence was observed for D. crawfordi at 12 and 30°C and for ^. ludens at 10 and 31°C. Loni (1997) studied the effects of temperature on the development of P. concolor. Adult emergence was greatest between 18-25°C, markedly reduced at 15 and 28°C, and zero at 13and33°C. Pemberton and Willard (1918) recorded diapause for D. tryoni and D. fullawayi. Darby and Kapp (1934) observed delayed emergence in several individuals of D. crawfordi. Clausen et al. (1965) reported diapause in D. longicaudata strains collected from areas having cool winters. Ashley et al. (1976) observed an increase in delayed emergence of D. longicaudata larvae at the lowest temperature (22°C) and likewise with low moisture concentration. Finally, Aluja et al. (submitted) recorded diapause in Mexican populations of D. areolatus, D. longicaudata and U. anastrephae, and also in Aganaspis pellenaroi (Brethes) and Odontosema anastrephae Borgmeier (Hymenoptera: Eulophidae).

PAGE 30

22 In conclusion, tenqjerature tolerances could determine the limits of distribution for fruit flies and their parasitoids. Laboratory studies of tenq)erature effects on Caribbean fruit fly parasitoids could help ascertain the influence of temperature on their distribution in Florida.

PAGE 31

CHAPTERS LARGE-SCALE DISTRIBUTION PATTERNS OF CARIBBEAN FRUIT FLY PARASITOIDS IN FLORIDA Studies from many geographical regions have indicated that distribution of fruit flies and their parasitoids may be affected by a variety of factors, including tenq)erature, precipitation, host fruit availability, and interspecific con^)etition (Chapter 2). Parasitoids may themselves be affected by host fly abimdance. In addition to mean or extreme temperature and precipitation, parasitoids may be influenced by the variance of these factors. In particular, the variability of abiotic factors among months could affect the temporal dynamics of hosts. Fruit yield is highly dependent on environmental factors, and can be adversely affected by tenqjeratures or rainfall that are periodically higher or lower than optimal (Petr 1991, Raper and Kramer 1983). High variance in these abiotic factors would lead to greater variability in the temporal availability of hosts. Parasitoid species may differ in their ability to survive through periods of low host abundance. Three species of parasitoids are commonly recovered from the Caribbean fruit fly, Anastrepha suspense (Loew), in Florida: Diachasmimorpha longicaudata (Ashmead), Doryctobracon areolatus (SzepUgeti) and Utetes anastrephae (Viereck) (Hymenoptera: Braconidae: Opiine). Studies conducted at several locations in Florida suggest that distribution of parasitoid species may differ at various sites (see Chapter 1). 23

PAGE 32

24 The objective of this chapter was to determine the parasitoid distribution throughout central and southern Florida, and identify abiotic and biotic fiictors possibly influencing this distribution. Materials and Methods Fruit Sampling Host finits of A. suspensa were collected in 23 towns in central and southern Florida (Figure 3-1). Sample sites were chosen to represent various regions of the peninsula. Thus samples were collected from 5 sites along the Atlantic coast from Melbourne (28.1° N) to Miami (25.8° N), 7 sites along the Gulf of Mexico coast from Tampa (28.0° N) to Naples (26.1° N), and 1 1 sites in the interior from Dade City (28.4° N) to Belle Glade (26.7° N) and LaBelle (26.8° N). Interior sites were situated along various north-south routes, e.g., US 17 (Lakeland, Wauchula and Arcadia) and US 27 (Haines City, Lake Wales, Lake Placid and Belle Glade). Clewiston, on the south-west coast of Lake Okeechobee, and Immokalee, 37 km south of LaBelle, were not sampled because of previous mass releases of D. longicaudata at these locations (Sivinski et al. 1996). Note that there are no interior towns south of those indicated due to the presence of the Everglades. Sanq)ling was not always limited to the town indicated, and often included collections in adjacent towns. Many Melbourne samples were actually collected in southern Brevard County, and St. Petersburg samples were collected throughout Pinellas County. Lake Placid samples include some from Sebring, and Punta Gorda includes

PAGE 33

Figure 3-1. Fruit sample collection sites.

PAGE 34

26 samples from Port Charlotte. On the other hand, two large coimties had more than one collection site. Haines City, Lakeland and Lake Wales are all in Polk County, while both West Palm Beach and Belle Glade are in Palm Beach County. Samples were collected in August 1994, and monthly from January-September 1995. Additional monthly samples were collected at Melbourne, Bradenton, Venice and Okeechobee from March-May 1996 and at St. Petersburg in May 1996. Sampling was not performed from October through December, because at the majority of sites primary host fruits are uncommon during this period in most years (Tim Holler, pers. comm.). All samples were collected within a single week each month. Sanpling at Ft. Pierce was conducted by Tim HoUer, USDA-APHIS-PPQ, in March and May 1993, prior to augmentative releases of D. longicaudata. Sampling at most sites during 1994 and 1995 was performed by the Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Bureau of Plant Inspection. Sampling at other sites during 1994 was by USDAAPHIS. Fruits sampled included loquat (Eriobotrya japonica (Thimb.)), Surinam cherry {Eugenia imiflora L.), Cattley guava {Psidium cattleianum Sabine) and conunon guava (Psidium guajava L.). Loquat samples were collected from January-April, most Surinam cherry samples from April-June, most Cattley guava from July-August, and most common guava from August-September. Note that additional fruiting periods may occur, especially at southern sites. Up to 12 samples were collected for each site during a single montL Each sample included fruits collected from a single tree. Fruits collected were ripe and usually without holes caused by beetles or exiting larvae. They were preferably collected from the tree, but occasionally supplemented with fruits from the ground.

PAGE 35

27 Numbers of samples and total fiiiits collected at various sites are detailed in Table 3-1. Total sample numbers varied widely among sites, ranging from 17 samples collected at Ft. Pierce to 102 at LaBelle. The nimiber of samples collected was dependent primarily on availability of host fruit. Note that the efficiency of fruit sampling may have varied among sites, as it was often conducted by different personnel at varioiis sites. Cattley guava was the least commonly collected host, with 83 samples, compared with 549, 377 and 290 samples of loquat, Surinam cherry and common guava, respectively. Fruits were placed within a bucket upon a metal screen. The bucket had holes for ventilatioiL It was covered with a plastic lid and its inside was lined with cloth to prevent entry of insects after fruit collection and escape of insects emerging from the fiiiit sample. Moist fine vermiculite (ca. 15 ml water per 100 cm^ vermiculite) was placed at the bottom of the bucket. Mature fruit fly larvae exited the fruit and pupated in the vermiculite. At the end of each sampling week, buckets were collected from various locations and transported to Gainesville. Buckets were maintained at 25.5° C, except in 1995 when they were kept in a warehouse at ambient temperatures. Puparia were sifted from the vermiculite 13-15 days after fiaiit collection, and transferred to 250 ml plastic containers. These containers were initially covered with a solid lid, which was replaced after ca. one week with a screened Ud. This was done to assure that the vermiculite did not dry out, but was also not so moist as to allow development of fimgi. Containers were maintained at 25.5° C and ambient humidity.

PAGE 36

Table 3-1. Total number of host fruit samples and total fruits collected at various sites. Site Loquat Surinam cherry Cattleyguava Common guava Samples Fruits Samples Fruits Samples Fruits Samples Fruits Arcadia 35 1162 11 1076 3 94 10 51 Belle Glade 8 139 31 1202 1 22 31 100 Bradenton 32 873 24 1433 1 1 7 59 Dade City 29 1017 0 0 0 0 6 13 Ft. Lauderdale 9 321 30 1469 2 50 13 47 Ft. Myers 27 854 14 1090 19 510 5 50 Ft. Pierce 6 610 10 1550 0 0 1 22 Haines City 22 567 9 349 0 0 0 0 LaBelle 39 1272 28 2021 5 210 30 226 Lakeland 35 663 16 614 4 100 26 122 Lake Placid 37 1261 19 1534 3 98 11 75 Lake Wales 21 713 18 916 0 0 29 192 Melbourne 42 1241 21 1181 4 140 17 166 Miami 3 72 12 553 2 27 31 171 Naples 26 630 23 1129 24 806 3 Okeechobee 21 713 20 1385 0 0 5 22 Punta Gorda 25 645 19 1160 10 353 0 0 St. Cloud 12 351 16 621 0 0 21 191 St. Petersburg 29 685 11 736 0 0 10 28 Tampa 29 561 0 0 0 0 8 Venice 15 441 20 1615 0 0 0 0 Wauchula 40 1494 15 1185 2 61 22 133 W.Pahn Beach 7 148 10 345 3 41 5 22

PAGE 37

29 Flies and parasitoids emerged within the containers, and were counted when no more Uve insects were observed. Parasitism levels for each species were calculated for each sample as the ratio between the number of parasitoids of the relevant species emerging and the sum of all flies and parasitoids emerging. This assumes that neither the flies nor the parasitoids diapause, and that mortality levels of the fly pupae and immature parasitoids are similar. In Florida, emergence rates of pupae held indoors under controlled temperatiire and himiidity are typically ca. 90% (Sivinski, pers. comm.), so that it is unlikely that a significant proportion of parasitoids undergo diapause. In contrast, diapause appears to be quite prevalent among parasitoids held under seminatural conditions in Mexico (Aluja et al., submitted). Note that true levels of parasitism are underestimated, because finit removed from the field include host eggs and larvae that may have been parasitized if left in place. However, comparisons of parasitism levels as measured should reflect relative abundance. Abiotic Environmental Data Temperature and precipitation data were obtained from the Southeast Regional Climate Center, Columbia, South CaroUna. Weather stations exist m most of the towns included in the study. However, there are no data for the vicinity of Dade City and Haines City. Data from Winter Haven and Avon Park were used to represent Lake Wales and Lake Placid-Sebring, respectively. The following variables were obtained: mean annual temperature, mean minimimi temperature for the coldest month of the year, mean maximum ten:q)eratijre for the warmest month of the year, extreme annual minimimi and maximum ten^ratures, and aimual precipitation (Table 3-2).

PAGE 38

30 Table 3-2. Mean precipitation and temperature values for towns in central and southern Florida, for the years 1980-1996. Data were obtained from the Southeast Regional Climate Center, Columbia, South Carolina. Town Annual Mean precipitation annual (mm) temp (°C) Mean minimimi temp ("C)" Extreme minimum temp (°C) Mean maximum temp ("C)" Extreme maximum temp CQ Arcadia 1341 22.2 8.0 -3.0 33.4 36.5 Avon Park 1263 22.3 7.6 -3.0 33.4 36.1 Belle Glade 1290 22.8 9.6 -0.8 33.3 35.5 Bradenton 1412 22.8 9.0 -1.8 33.6 36.0 Ft. Lauderdale 1672 24.4 13.9 2.9 32.3 35.3 Ft. Myers 1390 23.9 11.3 0.7 33.9 36.6 Ft. Pierce 1403 23.0 9.3 -1.8 33.4 36.5 LaBelle 1310 23.3 9.0 -1.6 34.1 36.8 Lakeland 1320 23.1 9.1 -3.2 34.6 37.2 Melbourne 1243 22.5 9.2 -1.5 32.5 36.0 Miami 1509 24.8 14.3 3.7 32.8 35.8 Naples 1325 23.8 11.3 0.6 33.5 35.9 Okeechobee 1224 23.1 8.3 -0.9 30.5 36.7 Punta Gorda 1309 23.5 .; 10.2 -0.9 33.8 36.1 St. Cloud 1268 22.6 8.6 -2.6 33.4 36.0 oi. reiersDurg 1270 23.3 11. 3 1.1 32.8 35.8 Tampa 1133 22.7 9.2 -1.8 33.0 35.5 Venice 1140 22.9 8.8 -0.9 33.4 35.4 Wauchula 1300 22.8 9.7 -2.6 34.0 36.4 W. Palm Beach 1609 24.1 12.8 1.8 32.7 35.5 Winter Haven 1286 23.0 9.2 -2.2 34.0 36.7 "For the coldest month of the year. ^or the warmest month of the year. Temp = temperature.

PAGE 39

31 The annual variance of the following monthly values were calculated: mean temperature, mean minimum temperature, mean maximum temperature, extreme minimimi tenq)erature, extreme maximum ten^rature and precipitatioiL The mean values of these variables for the years 1980-1996 were used for analysis. Host Fly and Host Plant Data A. suspensa catch data from McPhaU traps were obtained from the North Fbrida Research and Education Center, Quincy, Florida. Trapping was performed by the U.S. Department of Agriculture and by the Florida Department of Agriculture and Consumer Services, Division of Plant Industry. Only trappii^ data from urban locations were used, in order to conform with parasitoid data, which were also from urban sites. Data were available in the form of monthly numbers of flies per trap for various counties, and exact identity of the town(s) where fraps were situated was unknown. Following are the coimties for which data were available, and in parentheses the fruit collection site to which they were related in subsequent analyses: Brevard (Melbourne), Broward (Ft. Lauderdale), Charlotte (Punta Gorda), Collier (Naples), Dade (Miami), De Soto (Arcadia), Hardee (Wauchula), Highlands (Lake Placid), Hillsborough (Tampa), Lee (Ft. Myers), Manatee (Bradenton), Okeechobee (Okeechobee), Pahn Beach (Belle Glade), Pinellas (St. Petersburg), Polk (Lakeland), Sarasota (Venice), St. Lucie (Ft. Pierce). Trapping data were for the years 1992-1996, except for Brevard, Broward, Dade, Pahn Beach and Pinellas Counties, for which data were for 1990-1996. Variables used for analysis included mean monthly catch, minimal monthly catch, and maximal monthly catch. These were calculated for each year, and the mean annual value for each variable (Table 3-3) used for analysis.

PAGE 40

32 Table 3-3. Numbers of Caribbean firuit flies captured in McPhail traps for various counties in central and southern Florida. Traps were maintained by the Florida Division of Plant Industry and the U. S. Department of Agriculture. Data were obtained from the North Florida Research and Education Center, Quincy, Florida. Countv rcorresnondinp stte^ IVlliilillUlll ivioAir 1 d A /.o n fil I 1 < < I I J.J rharlotte rPiinta GorHfl"^ to U.J 1 Z /O. / n 0 1U4.6 1 m l.U/ 1 OQ /I Dp Sntn ( AroaHifi^ J. 1 U.J J 1 O f» X XcU vL^^ w dUviiuia 1 u.io Jin u.oo JJ.O X xxLLo LAji yj Ki^lx y X dXiilla j J.J U.jO Zo.U T,pe CPi A/fvpr^^ ^4 4 Z.Zl 14/. / ivianaiee (^rjraaentonj 3.0 0.15 15.5 Okeechobee (Okeechobee) 1.1 0.02 6.1 Palm Beach (Belle Glade) 8.8 0.14 46.2 Pinellas (St. Petersburg) 4.0 0.03 22.6 Polk (Lakeland) 11.8 0.72 52.3 Sarasota (Venice) 19.7 1.22 97.2 St. Lucie (Ft. Pierce) 13.4 1.09 42.9 Mean values for the years 1990-1996 for Brevard, Broward, Dade, Pabn Beach and Pinellas counties, and 1992-1996 for other counties. "Mean monthly catch. 'Minimum monthly catch. 'Maximum monthly catch.

PAGE 41

33 A siirvey was performed to determine the density of host fruit trees in various towns. Quadrants were chosen from various regions of each town. Previous experience suggested that older middle-class neighborhoods were the best areas for A. suspensa hosts. Specific quadrants were picked that appeared on a map to match this designation. Upon arrival, obviously unsuitable quadrants were dismissed, and others chosen from the map to replace them Four quadrants were sampled in each town, except LaBelle where 5 quadrants were sanqjled. Host trees were counted during a slow drive through the neighborhood. Thus trees in back yards were counted only if observed from the street. As towns differ in the size and visibility of backyard properties, the number of trees present but not observed per unit area would presumably also differ. Therefore, comparisons of towns based on number of trees counted per imit area may not be reUable. Rather, relative density was estimated as the number of trees observed per km of road. Distances traveled per quadrant ranged from 4.3-14.8 km, but were usually between 5-10 km. Statistical Analysis Parasitism data were subjected to an arcsine square root transformation before analysis. All analyses were performed using SAS statistical software. Environmental factors could be associated with either (1) absolute parasitoid distribution, i.e., presence or absence at various sites, or (2) relative abundance of parasitoids among sites in which they are present, as measured by parasitism level. The factors associated with each response may differ. Associations of environmental fectors with presence or absence of each parasitoid species were analyzed using logistic regression models (SAS Institute 1989). All

PAGE 42

34 tenq)erature, precipitation and fruit tree density factors were tested separately in these analyses. Associations of ten^erature, precipitation and host fruit tree density with parasitism levels were analyzed using linear regression models. Models examining parasitism of all species combined included all sites sampled. Models examining parasitism levels of each species separately included only sites in which the relevant parasitoid species was collected. The various fruit types were analyzed separately. Initially, all temperature, precipitation and fruit tree density factors were included, and their relative fit with the model determined by the forward selection procedure (SAS Institute 1989). Ultimately, only the factor best explainmg parasitoid abimdance remained in the final single linear regression model. In addition to single regression, multiple regression models were examined including all factors significantly related with parasitism levels. Host fly population levels are not mdependent of the previously described fectors, i.e., temperature, precipitation and host fruit tree density. Therefore, host fly data could not be included as factors in the previous analyses. Separate Unear regression models were examined relating parasitism levels with fly trapping variables. All sites were included in this analysis. Thus in this case I did not differentiate between presence or absence of parasitoid species and their relative abimdance.

PAGE 43

35 Results Distribution and Abundance of Parasitoids Numbers of samples containing parasitoids and numbers of parasitoids emerging for various towns and host finits are detailed in the Appendix and summarized in Tables 3-4 and 3-5. With data from all towns combined, parasitism levels were higher in Surinam cherry and Cattley guava than in loquat or common guava for both D. areolatus and D. longicaudata. For U. anastrephae, parasitism levels were higher in Surinam cherry than m any other fruit (Table 3-6). Parasitism by U. anastrephae was extremely low in common guava, with only 8 individuals recovered from 4 samples (Table 3-5). These results are consistent with the findings of Sivinski (1991; Sivinski et al. 1997) that smaller fruits have higher levels of parasitism. As expected, the differences in parasitism levels among fruit types is largest for U. anastrephae, which has a relatively short ovipositor, and thus less access to larvae deep within large fruits. Overall, D. areolatus was more abundant than D. longicaudata in Surinam cherry and common guava, but not in loquat or Cattley guava (Table 3-6). Both species were more common than U. anastrephae in loquat and common guava, but mean parasitism levels of Z). longicaudata and U. anastrephae were not significantly different in Surinam cherry or Cattley guava. With all data combined, it appears that parasitism levels are quite low (Table 3-6). Note, however, that this includes data from sites where certain parasitoid species were totally absent from samples. Even where parasitoids were recovered, many samples did

PAGE 44

36 Table 3-4. Numbers of samples collected and insects emerging for various sites. Each sample includes fruits from a single host tree. Site Number of samples Niraiber of insects emerging Total With CFF* With Da" With Dl" With Ua" CFF* Da" 01*= Ua** Arcadia 59 57 27 0 1 3095 589 0 3 BeUe Glade 71 63 0 20 9 3462 0 106 27 Bradenton 64 48 0 1 4 1816 0 6 33 Dade City 35 32 0 0 0 672 0 0 0 Ft. Lauderdale 54 48 0 12 9 2746 0 214 59 Ft. Myers 65 59 9 22 19 3092 27 156 79 Ft. Pierce 17 17 0 7 8 1352 0 10 110 Haines City 31 18 1 0 0 281 1 0 0 LaBelle 102 95 42 43 2 3267 768 482 7 Lakeland 81 63 13 0 1 3629 172 0 2 Lake Placid 70 64 23 0 1 2837 356 0 1 Lake Wales 67 59 19 0 0 2500 230 0 0 Melbourne 84 53 0 0 0 961 0 0 0 Miami 48 40 0 11 4 2539 0 81 10 Naples 76 49 2 9 6 1210 5 58 36 Okeechobee 46 40 6 4 1 1748 181 9 7 Punta Gorda 54 36 5 1 3 1479 19 1 11 St. Cloud 49 37 0 0 2 1288 0 0 16 St. Petersburg 50 34 0 0 0 1354 0 0 0 Tampa 37 28 2 0 0 1198 2 0 0 Venice 35 31 1 0 13 2097 1 0 60 Wauchula 79 75 36 0 3 4368 599 0 16 W. Pahn Beach 25 19 0 3 1 439 0 51 5 'CFF = Caribbean fruit fly. "Da = Doryctobracon areolatus. "Dl = Diachasmimorpha longicaudata. **Ua = Utetes anastrephae.

PAGE 45

37 Table 3-5. Numbers of samples collected and parasitoids emerging for various host fruits. Each sample includes fruits from a single host tree. Host Number of samples Number of insects emerging it Total With With With With CFF Da" Dl'= Ua** CFP Da" Dl" Ua*^ Loquat 549 418 56 30 13 13618 657 418 63 Surinam cherry 377 326 64 55 63 12329 1197 410 377 Cattley guava 83 60 14 15 7 2147 280 97 34 Common guava 290 261 52 33 4 20395 816 249 8 "CFF = Caribbean fruit fly. "Da = Doryctobracon areolatus. *T)1 = Diachasmimorpha longicaudata. *'Ua = Utetes anastrephae. Table 3-6. Mean percent parasitism (SE) for various host fruits. Fruit D. areolatus D. longicaudata U. anastrephae Loquat 2.0 (0.3) B a 1. 5 (0.4) B a 0.2(0.1) Bb Surinam cherry 7.8(1.1) A a 3. 6 (0.6) A b 3.0 (0.6) A b Cattley guava 7.1 (2.8) A a 5. 3 (1.6) A ab 1.1 (0.5) B b Common guava 2.5 (0.4) B a 1. 0(0.3) B b 0.03 (0.02) B c Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio=100.

PAGE 46

• ' 38 not contain parasitoids (Tables 3-4 and 3-5). When maximum parasitism is considered, it becomes apparent that all three parasitoid species are capable of achieving high levels of parasitism. Over 50% parasitism was observed in certain samples at 6 sites for D. areolatus, at 5 sites for D. longicaudata, and at 2 sites for U. anastrephae (Table 3-7). Additionally, as mentioned above, these observations are almost certainly underestimates of the true parasitism levels. Mean levels of parasitism for the braconid parasitoids at various sites are summarized in Tables 3-8 through 3-11. At three northern locations, Dade City, Melbourne and St. Petersburg, no parasitoids were found (Table 3-4). D. areolatus was absent jfrom the Atlantic coast, and also was not collected at Belle Glade, St. Cloud or Bradenton (Figure 3-2). It was most common at interior locations, and relatively rare along the Gulf coast (Figure 3-3). However, distance from the coast was not a significant predictor of D. areolatus abundance, perhaps because of its absence at two interior locations. Highest mean parasitism levels observed for various fruits were 6.6% in loquat at Arcadia, 35.8% in Surinam cherry at LaBelle, 79.7% in Cattley guava at Arcadia, and 10.2% in common guava at Arcadia. D. longicaudata was not collected at interior locations north and west of Lake Okeechobee, or at the most northern locations along both coasts (Figure 3-2). It was uncommon at all locations at the northern end of its distribution, with the exception of LaBelle (Figure 3-4). D. longicaudata abundance was significantly greater at lower latitudes in both Surinam cherry and common guava (F=12.9, p=0.002 and F=6.5, p=0.02, respectively). Highest mean parasitism levels observed for various fruits were

PAGE 47

39 13 c o E E o U O > o o lO'*! I— ooloo«o!csi r5 ! 0\ 0^ 00 ! I I O I I O CO 1 I >/-> — SO I OS ro\ — H . ^ ^ d «^ o o sd I r4 I I o so r«-) i — w — . >o : O fS I S0fNr<^O(NO — OO-— r— fS'^or~oooo>ofn'-«ooo cn — OS — rn d — so O t-^ o I O^^OO00(^^0^00OOS0S0T}OO O so so O 00 — — _ O O O *-»--•*-• l2 ^ ^ d d c 3 o u X) o E J u — 4> CI] o. w S S 2 O ^ s O CO ^ § ^ es — « ^ U 0. 3 *J *J ft. (/3 ai C8 OQ 3 E a. o o C •£ 3 ft. Ills I t c s I i I s: •o a 5> as Q

PAGE 48

40 Table 3-8. Mean percent parasitism (SE) in loquat for various sites. Site D. areolatus D. longicaudata U. anastrephae Arcadia 6.6 (1.8) A a 0 C b 0 C b Belle Glade 0 B 0 c 0 c Bradenton 0 B 0 c 0.1 (0.1) c Dade City 0 B 0 c 0 c Ft. Lauderdale 0 B 2.0 (2.0) BC 0.7 (0.7) c Ft. Myers 0.2 (0.1) AB b 6.9(1.1) AB a 1.1 (0.5) BC ab Ft. Pierce 0 B 0.7 (0.6) BC 4.6 (3.3) A Haines City 0 B 0 c 0 c LaBelle 4.9(1.5) AB b 10.4 (0.3) A a 0 c Lakeland 0 B 0 C 0 c Lake Placid 3.1 (1.9) AB a 0 c b 0 c U u Lake Wales 2.7 (1.4) AB a 0 c b 0 c u D Melbourne 0 B 0 c 0 c Miami 0 B 1.8(1.8) BC 1.8(1.8) B Naples 0 B 0.05 (0.05) C 0.4 (0.4) c Okeechobee 0 B 0 c 0 c Pvmta Gorda 0.3 (0.3) AB 0 c 0 c St. Cloud 0 'B' ; 0 c 0 c St. Petersburg 0 B 0 c 0 c Tampa 0.05 (0.05) B 0 c 0 c Venice 0 B 0 c 0.06 (0.06) C Wauchula 6.5 (1.5) AB a 0 c b 0 c b W. Palm Beach 0 B 7.2 (7.2) AB 0 c Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio=100.

PAGE 49

41 Table 3-9. Mean percent parasitism (SE) in Surinam cherry for various sites. Site D. areolatus D. longicaudata U. anastrephae Arcadia 25.1 (11.1) ABC a 0 B b 0.4 (0.3) A b Belle Glade 0 E b 12.1 (3.5) A a 7 9 G 3) ' V-'--'/ A a Bradenton 0 E 0.6 (0.6) B 2.4 (1.6) A Ft. Lauderdale 0 E a 7.3 (3.2) AB a 5\(21') J. I 1 ) A a Ft. Myers 0.6 (0.6) E b 3.6 (2.4) B ab 8.4 (2.8) A a Ft. Pierce 0 E b 0.7 (0.4) B ab etc/ 6.8 (3.4) A a Haines City 12.5 (12.5) CDE 0 B 0 A LaBelle 35.8 (6.3) A a 13.2 (3.5) A b 1.4(1.1) A c Lakeland 4.6 (2.2) E a 0 B b 0 A b Lake Placid 20.2 (6.9) BCD a 0 B b 0.07 (0.07) A b Lake Wales 10.1 (5.5) DE a 0 B b 0 A b Melbourne 0 E 0 B 0 A Miami 0 E b 15.4 (7.2) A a 1.8(1.5) A b Naples 0.4 (0.4) E 2.7 (2.1) B 7.3 (6.6) A Okeechobee 10.9 (5.1) DE a 0.8 (0.7) B b 0.6 (0.6) A b Punta Gorda 3.5 (2.8) E 0 B 2.4(1.7) A St. Cloud 0 E 0 B 3.8 (2.7) A St. Petersburg 0 E 0 B 0 A Venice 0.06 (0.06) E b 0 B b 4.6 (2.0) A a Wauchula 32.4 (8.9) AB a 0 B b 2.0(1.3) A b W. Palm Beach 0 E 3.5 (2.8) B 1.7(1.7) A Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio=100.

PAGE 50

• ^ 42 Table 3-10. Mean percent parasitism (SE) in Cattley guava for various sites. Site D. areolatus D. longicaudata U. anastrephae Arcadia 79.7 (10.2) A 0 0 A Ft. Lauderdale 0 D 0 0 A Ft. Myers 0.5 (0.3) D b 8.8 (3.4) a 1.2 (0.8) A b LaBelle 17.9(11.9) C 5.7 (3.6) 0 A Lakeland 3.6 (3.6) D 0 0 A Lake Placid 65.8 B 0 0 A Melbourne 0 D 0 0 A Miami 0 D 0 10.0 (10.0) A Naples 0.6 (0.6) D 9.8 (5.3) 1.8(1.4) A Punta Gorda 0.3 (0.3) D 0.3 (0.3) 0 A W. Palm Beach 0 D 0 0 A Means within a colimm followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio=100. 10.4% in loquat at LaBelle, 15.4% in Surinam cherry at Miami, 9.8% in Cattley guava at Naples, and 9.6% in common guava at Ft. Myers. U. anastrephae was widespread, having been collected at most locations (Table 34). Parasitism levels for this species were relatively low, especially at most interior locations (Figure 3-5). Highest mean parasitism levels observed for various fiaiits were 4,6% in loquat at Ft. Pierce, 8.4% in Surinam cherry at Ft. Myers, 10.0% in Cattley guava at Miami, and 1 .0% in common guava at Ft. Myers.

PAGE 51

43 Table 3-11. Mean percent parasitism (SE) in common guava for various sites. Site D. areolatus D. longicaudata U. anastrephae Arcadia 10.2(4.1) A a 0 B b 0 B b Belle Glade 0 B 2.0 (1.4) B 0 B Bradenton 0 B 0 B 0 B Dade City 0 B 0 B 0 B Ft. Lauderdale 0 B b 3 5 0 7) B 0.2(0.2) B b Ft. Myers 2.4(2.1) AB 9 6 r4 4) A 1.0(0.7) A Ft. Pierce 0 B 0.4 B 0 B LaBelle 7.1 (2.8) AB a 3.0(1.2) B ab 0 B b Lakeland 2.5 (1.4) AB a 0 B b 0.03 (0.03) B ab Lake Placid 4.9 (2.0) AB a 0 B b 0 B b Lake Wales 4.5 (1.8) AB a 0 B b 0 B b Melbourne 0 B 0 B 0 B Miami 0 B b 1.3 (0.7) B a 0 B b Naples 0 B 4 9 f4 9^) AB 0 B Okeechobee 0 B 0 8 CO n 0 B St. Cloud 0 B 0 B 0 B St. Petersburg 0 B 0 B 0 B Tampa 0.05 (0.05) B 0 B 0 B Wauchula 6.2 (2.9) AB a 0 B b 0 B b W. Palm Beach 0 B 0 B 0 B Means within a column followed by the same upper-case letter, and means within a row followed by the same lower-case letter, are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio==100.

PAGE 52

44 The distribution ranges of D. areolatus and D. longicaudata overlap only within a limited area from Lake Okeechobee to the Gulf of Mexico coast (Figure 3-2). In feet, only at LaBelle are both common (Figure 3-6). Interestingly, LaBelle was among the sites with the highest parasitism levels for both D. areolatus and D. longicaudata. With all sites included, there was no significant relationship between parasitism levels of D. areolatus and D. longicaudata in Surinam cherry (Spearman correlation coefficient = -0.27, p= 0.23). However, with LaBelle excluded, parasitism by the two species was negatively related (Spearman correlation coefficient = -0.48, p=0.034). Temporal and spatial dynamics at LaBelle are explored in Chapter 4. Although U. anastrephae is widespread, it is most common at coastal locations and at Belle Glade, i.e., at locations where D. areolatus is absent or rare and where D. longicaudata is commoa With all sites included, parasitism levels of U. anastrephae and D. longicaudata in Surinam cherry were positively related (Spearman correlation coefficient = 0.53, p=0.014), indicating a similar distribution pattern for both species. Interestingly, when considering only sites where both species were recovered, no significant relationship was observed (Spearman correlation coefficient = -0.02, p=0.96). This suggests that these species do not unpact each other on a local level. With all sites included, no significant relationship was observed between parasitism levels of U. anastrephae and D. areolatus in Surinam cherry (Spearman correlation coefficient = -0.35, p=0.12). However, when only sites with both species present were considered, parasitism was negatively related (Spearman correlation coefficient = -0.70, p=0.044). This suggests that U. anastrephae and D. areolatus may have a negative impact on each other. Only at Punta Gorda are these species found

PAGE 53

45 together in similar numbers (Figure 3-6). Note that at this location parasitism by both species was low, with only 19 D. areolatus and 1 1 U. anastrephae recovered (Table 3-4). Mean parasitism at Punta Gorda in Surinam cherry was only 3.5% and 2.4% for D. areolatus and U. anastrephae, respectively (Table 3-9). Thus significant con:q)etition at this site was unlikely. A fourth parasitoid, the eulophid Aceratoneuromyia indica Silvestri, emerged from two puparia collected from guava at Belle Glade in February 1995. To my knowledge, this is the only report of this species outside the area of its introduction in Dade County, ca. 140 km to the south. Host Plant Density Sample sites varied considerably in the relative density of host fruit trees (Table 3-12). Loquats are most abundant at Dade City, Lake Wales, Melbourne and Tan^a, all of which are situated at high latitudes (Figure 3-1). At Dade City, the highest latitude site in this study, other hosts are rare, with only one small Surinam cherry hedge and one common guava located. Lowest numbers of loquats were found at Ft. Myers, Miami, Ft. Lauderdale and Belle Glade. Surinam cherries are most abundant at Miami, St. Petersburg and Belle Glade, Cattley guavas at Naples, Ft. Pierce and LaBelle, and common guavas at LaBelle. Note that at southern coastal towns Uke Miami, additional tropical host trees occur, which are not included in this survey. The most important of these hosts is tropical ahnond, Terminalia catappa L.

PAGE 54

46 Figure 3-2. Distribution of Doryctobracon areolatus and Diachasmimorpha longicaudata. Includes data from the current study and Sivinski et al. (1996).

PAGE 55

Figxu-e 3-3. Parasitism by Doryctobracon areolatus in Surinam cherry.

PAGE 56

48 N
PAGE 57

/ Figtire 3-5. Parasitism by Utetes anastrephae in Surinam cherry.

PAGE 58

50

PAGE 59

51 Table 3-12. Mean host fruit plant density (trees/ km of road) (SE) in various towns. 1 UWIl ix>4uai Surinam cherry Cattley guava Common guava Arcaaia 3.4 (yA) Cde 2.6 (0.4) ei A O /"A 1 \ 1_ J 0.8 (0.1) bed 0.10 (0.06) b £>eue vjiaae 1.2, (U.j) ei 1 1 if\ o\ 1-.— 7.7 (0.8) be A /A A'^X J 0.2 (0.02) cd 0.5 (0.3) b Bradenton i.i (O.o) cue 4.6 (0.7) cde 0.3 (0.1) cd 0.02 (0.02) b uaae uity 0.3 (U.4) a 0.04 (0.04) I A J 0 d 0.04 (0.04) b ri. i/auaeroaie 1 o /n /i\ 1 .0 ei C 'J /A A\ 5.3 (0.9) bcde 0.5 (0.2) bed Ob ri. jviyers A Q /^A T\ -f u.y (u.z) I 2.7 (0.3) ei 0.6 (0.1) bed 0.06 (0.04) b ri. rierce 4.4 (U.D) abed A n /A A\ «.J. 4.9 (0.9) cde 1.4 (0.7) b 0.2 (0.1) b jL>ac>eue 1 A /'A 0\ j.u (U.oj caei O /A 0\ ^ 3.4 (0.8) e 1 A /A 1_ 1.0 (0.2) be 2.4 (0.9) a L,aKClaHQ J.J cde 3.7 (0.4) e A J 0 d 0.06 (0.04) b ivaKe r laciG z.y (1.2) cdei A A /A A\ J 4.9 (0.9) cde 0.5 (0.2) bed 0.5 (0.3) b Ijokc waies A 1 /'A <;\ o. 1 (U.o) ao 7.4 (0.8) bed 0.5 (0.2) bed 0.4 (0.2) b Melbourne 5.y (1.1) ab 3.2 (1,0) ef 0.3 (0.1) cd 0.3 (0.1) b ivxiciini 1 1 /"A /1^ -T 1.1 (.U.4J I 1 O A A\ — 12.9 (3.9) a 0.3 (0.2) cd 0.7 (0.3) b z.D (^u. J J del A O /A 'i\ ^ 4.6 (U.3) Cde 3.9 (1.2) a 0.10 (0.06) b W&.CCC no Dec 1 1 /"A ox ^A,^f 5.1 \yj.2.) cdei 1 T /A y4\ 3.7 (0.4) e 0.5 (0.1) cd 0.6 (0.3) b r unxa uoroa "5 A /I TX U>^J» 3.9 (1.7) bcde 4.2 (1.1) de 0.3 (0.1) cd Ob St. Cloud 3.4 (0.6) cde 2.3 (0.6) ef 0.2 (0.1) cd 0.2 (0.1) b St. Petersburg 3.0 (0.2) cdef 8.2 (2.0) b 0.2 (0.1) cd 0.06 (0.03) b Tan^a 5.8 (1.1) ab 2.6 (0.5) ef 0.2 (0.1) cd 0.04 (0.04) b Venice 2.4 (0.3) def 2.3 (0.6) ef 0.6 (0.2) bed Ob Wauchula 4.8 (0.6) abc 3.5 (1.0) e 0.8 (0.2) bed 0.8 (0.2) b Means within a column followed by the same letter are not significantly different, p=0.05 according to the Waller-Duncan k-ratio t test, and k-ratio=100.

PAGE 60

52 Relationships With Environmental Factors The northern limit of D. longicaudata distribution closely fits the isotherm of 10.5° C January mean minimum temperature (Fig\ire 3-7). The number of frost days also fits the distribution somewhat (see map in Femald 1981), but to a lesser degree. Mean minimum temperature for the coldest month was not a significant factor in the logistic regression analysis (x^=3.51, p=0.061). Rather, presence of this parasitoid was best explained by low variability in monthly temperatures (Table 3-13). Four different variance factors were negatively related v^dth presence of D. longicaudata, including variance of extreme maximum, extreme minimum, mean maximimi and mean temperature. Other factors significantly related with presence of D. longicaudata included mean and extreme minimum temperatures (positive relationships) and abundance of loquat trees (negative relationship). In contrast with D. longicaudata, variance of temperatures had a positive relationship with presence of D. areolatus. In this case the significant factor associated with parasitoid distribution was variance of extreme minimum temperatures (Table 3-13). Presence of D. areolatus was also positively associated with extreme maximimi temperatures. Note that simimer temperatures are often greater at northern and inland sites than at southern and coastal locations (Table 3-2). Presence of U. anastrephae showed no significant relationships with abiotic fectors, but was negatively related with abundance of loquat trees. Density of guava trees was positively related v^th total parasitism in both loquat and Surinam cherry (Table 3-14). Other factors significantly associated with total parasitism include variance of extreme minimvmi temperatures (positive relationship in Cattley guava), extreme maximum temperature (positive relationship in common guava)

PAGE 61

53 Figure 3-7. Relationship between the northern observed limit of Diachasmimorpha longicaudata distribution and January mean minimum temperature. The isotherm was copied from Femald (1981), based on data from the years 1960-1979.

PAGE 62

54 and variance of precipitation (positive relationships in loquat and common guava, multiple fector models). D. areolatus was more common at sites with lower mean temperatures, in both Surinam cherry and Cattley guava (multiple and single factor models, respectively; Table 3-15). Similarly, parasitism levels in common guava were negatively associated with extreme minimum temperatures (multiple factor model). Variance of extreme maximum temperatures was negatively related with D. longicaudata abimdance in Surinam cherry (multiple factor model). Variance of precipitation was positively associated with D. areolatus abundance in Cattley and common guava, and with D. longicaudata abimdance in loquat (multiple &ctor models). Guava tree density was significantly associated with parasitism levels of both D. areolatus (positive relationship in Surinam cherry) and D. longicaudata (positive relationships in loquat and Surinam cherry, multiple factor models) (Table 3-15). Loquat tree density was negatively related with D. longicaudata parasitism levels in Surinam cherry and common guava (multiple factor models). Similarly, Surinam cherry tree density was negatively related with D. longicaudata abundance in common guava (multiple factor model). A highly significant positive relationship was observed between Surinam cherry tree density and parasitism levels of U. anastrephae in Cattley guava. Interestingly, there were no significant relationships between tree density and parasitism on the same host. Highly significant positive relationships were found between fly trapping and parasitism by D. longicaudata in loquat and common guava (minimum and mean monthly capture of host flies, respectively; Table 3-16). Similar, but less significant.

PAGE 63

55 Table 3-13. Environmental factors significantly associated with presence or absence of parasitoid species, according to logistic regression analysis. Parasitoid species Factor (relationship) P D. areolatus Var extreme min temp (+) 0.039 4.26 Extreme max temp (+) 0.045 4.00 D. longicatuiata Var extreme max temp (-) 0.016 5.79 Var extreme min temp (-) 0.017 5.69 Var mean max i temp (-) 0.028 4.82 Var mean temp (-) 0.030 4.70 Mean temp (+) 0.035 4.46 Extreme min temp (+) 0.038 4.28 Loqnat density (-) 0.039 4.25 U. anastrephae Loquat density (-) 0.016 5.78 Extreme min Xtvap = Extreme max ten^ = Extreme minimum Extreme maximum annual temperature, annual temperature. Loquat density = Mean density of loquat trees. Var extreme min ten^ = Variance of monthly extreme minimum temperatures. Var extreme max temp = Variance of monthly extreme maximimi ten5)eratures. Var mean ten^) = Variance of monthly mean ten:q)eratures. Var mean max tenq) = Variance of monthly mean maximum tenqieratures. relationships were observed between minimum capture and abundance of U. anastrephae in loquat and Surinam cherry. No significant relationships were found between fly trapping variables and parasitism by D. areolatus. Discussion Factors possibly affecting parasitoid distribution and abundance can be grouped mto three categories: (1) abiotic factors, e.g. temperature and precipitation; (2) host

PAGE 64

56 Table 3-14. Environmental fectors significantly associated with parasitism levels of all parasitoid species combined. Host fruit Single factor model* Multiple factor model Factor (relationship) Factors (relationship) Loquat Guava density (+) ** 0.34 Guava density (+) Var precipitation (+) * ** 0.61 Surinam cherry Guava density (+) *** 0.52 Cattley guava Var extreme min temp (+) * 0.36 Common guava Extreme max temp (+) * 0.21 Extreme max temp (+) * 0.42 Var precipitation (+) Linear regression with the factor which best explains variation in the data, '^lultiple Unear regression with factors having significant linear relationships with the relevant parasitism level. '* p<0.05; * p<0.01; *** p<0.001. Guava density = Mean density of common guava trees. Extreme max temp = Extreme annual maximum temperature. Var extreme min temp = Variance of the monthly extreme minimum temperatures. Var precipitation = Variance of the monthly precipitation. availability; and (3) competition with other species. These environmental effects are summarized in Figure 3-8. The northern limit of D. longicaudata distribution is closely related to the January mean minimum isotherm of 10.5°C, as derived fi-om Femald (1981). This suggests that vmter ten^ratures hmit the distribution of this species. This could be an important factor affecting tropical parasitoid species in warm temperate locaUties. Snowball and Lukin (1964) suggested that winter temperatures may limit the estabUshment of Fopim arisanus in Australia. Sivinski et al. (1998) reported a reduction in winter in the abundance of D. longicaudata relative to that of D. areolatus in calamundin at LaBelle. This suggests

PAGE 65

57 Table 3-15. Environmental fectors significantly associated with parasitism levels of the various parasitoid species, for towns in which the relevant parasitoid was collected. Species" Host fruit Single factor model'' Muhiple factor model*^ Factor (relationship) p"* Factors (relationship) p" Da Loquat Surinam cherry Guava density (+) * 0.50 Guava density (+) Mean temp (-) ** ** 0.81 Cattley guava Mean temp (-) " 0.90 Mean temp (-) Var precipitation (+) *** * 0.98 Common guava FyItpitip min tpTnf> (-\ Var precipitation (+) *** ** 0.84 Dl Loquat Guava density (+) Var precipitation (+) 0.68 Surinam cherry Var extreme max temp (-) Guava density (+) Loquat density (-) ** *** n 87 U.o / Cattley guava Common guava Loquat density (-) Sur cherry density (-) * * 0.75 Ua Loquat Surinam cherry Cattley guava Sur cherry density (+) ** 0.68 Da = Doryctobracon areolatus; D 1= Diachasmimorpha longicaudata; Ua = Utetes anastrephae. 'linear regression with the factor which best explains variation in the data. ''Muhiple linear regression with factors havmg significant linear relationships with the relevant parasitism level. p<0.05; ** p<0.01; *** p<0.001. Sur cherry density = Mean density of Surinam cherry trees. Guava density = Mean density of common guava trees. Loquat density = Mean density of loquat trees. Extreme min temp = Extreme annual minunum temperature. Mean temp = Mean annual temperature. Var extreme max temp = Variance of the monthly extreme maximum temperatures. Var precipitation = Variance of the monthly precipitation

PAGE 66

58 Table 3-16. Fly trapping variables significantly associated with mean parasitism levels. Parasitoid Host fruit Variable (relationship)' P' D. longicaudata Loquat Minimum (+) *** 0.64 Common guava Mean (+) *** 0.68 U. anastrephae Loquat Minimum (+) ** 0.38 Surinam cherry Minimum (+) * 0.30 'Linear regression with the fector which best explains variation in the data. Mean = Mean monthly capture (flies/trap). Minimum = Minimum monthly capture (flies/trap). "* p<0.05; ** p<0.01; *** p<0.001. some climatic effect unrelated to host availability. A similar conclusion was reached by Snowball (1966) regarding abundance of F. arisanus in southern Australia. Laboratory studies suggest that D. longicaudata may be less tolerant to low temperatures than its host A. suspensa (Chapter 5). Further studies are needed to determine whether its tolerance is lower than that of D. areolatus. Note that mean minimum ten^erature was not a significant fector in the logistic regression analysis. Rather, absence of D. longicaudata was best e3q)lained by high variability in temperatures (Table 3-13). The possible significance of this observation is discussed below. Parasitoid populations interact with the temporal and spatial distribution of their hosts, and the dynamics of both host fruits and fly populations may be important. The three towns from which parasitoids were not collected all have some type of low host availability. At Dade City, the most northern town in this study, all host plants except loquat are absent or rare (Table 3-12). Thus, hosts are available for only a short period of time in early spring. Note that host larvae are not uncommon in loquat at Dade City, with

PAGE 67

59 91% of samples collected producing flies (Table 3-4). Host fly densities in Brevard county (Melbourne) are among the lowest of all coimties included in this study (Table 33). Only 63% of samples collected at Melbourne produced flies (Table 3-4). Hosts are also relatively rare in Pinellas county (St. Petersburg) (Table 3-3), with only 68% of samples collected producing flies (Table 3-4). Note also that Pinellas coimty is a peninsula, separated form the east and south by Tampa Bay. This body of water may be an ecological barrier to parasitoids, making colonization difficult. Fluctuations in fruit availability would depend primarily on abiotic conditions including temperature and precipitation (Petr 1991, Raper and Kramer 1983). High variability in tenqjeratures could lead to greater heterogeneity in the tenqwral occurrence of fruit. Thus the negative relationships between the variances of several tenperature variables and occurrence of D. longicaudata (Table 3-13) may indicate that this parasitoid is dependent on a relatively constant supply of hosts. This hypothesis is supported by the highly significant relationship between mmibers of flies trapped and parasitism by this species in loquat and guava (Table 3-16). Spatial distribution of host fruits is dependent upon (1) the number of fruits per tree and (2) the number of trees per unit area. The number of fruits per tree depends on ; ' ' ^ * various factors including tree size and age, degree of shading, and horticultural practices such as pruning, watering and fertilization. Although these fectors may vary among towns, no data on this are available. In several cases, parasitism appears to be related to host tree density. Guava tree density accounts for 34 and 52% of the variation in total parasitism for loquat and Surinam cherry, respectively (Table 3-14). It is a significant factor in the abundance of

PAGE 68

60 both D. areolatus and D. longicaudata (Table 3-15). Guava is usually the last available major host before the onset of winter. Thus its abundance may be an important determinant of the size of the overwintering population, which in turn affects the abundance of parasitoids in loquat and Surinam cherry the following spring. In Hawaii, D. longicaudata is more common in orchards than in wild guava (Vargas et al. 1993). The authors suggest that two possible factors contributing to this observation may be high tree densities and abundance of rotting fruit in commercial guava orchards. The density of Surinam cherry trees accounts for 68% of the variation in U. anastrephae abundance in Cattley guava (Table 3-15). This result is expected because Surinam cherry is the major host for U. anastrephae (Table 3-6), and it immediately precedes Cattley guava in fruiting. Note, however, that U. anastrephae was recovered from Cattley guava in three towns only (Table 3-10), and a regression analysis based on three points should be treated with caution. Significant negative relationships between fruit tree density and parasitism are probably artifacts. The absence of U. anastrephae in towns with high incidence of loquats may reflect a tendency of loquats to survive in towns that have an unsuitable climate for other tropical fruits, and that the lack of the latter is what actually accounts for the absence of parasitoids (loquats flourish and fruit well north of the normal range of ^. suspensa, pers. obs.). Dade City, Melbourne and Tampa, among the towns with the highest loquat densities, all have relatively low densities of other host fruits (Table 3-12). When all towns except LaBelle are considered, abundance of D. areolatus is negatively related to that D. longicaudata (Figures 3-3, 3-4 and 3-6). Diachasmimorpha

PAGE 69

61 longicaudata and D. areolatus are of similar size, and both have long ovipositors (Sivinski et al. 1997). They also show similar preferences to the host fruits (Table 3-6), and attack the same stage larvae (Lawrence et al. 1976, Chapter 6). Thus the potential for competition between these species is obvious. Note that the interaction between them is a new association, as D. areolatus is a neotropical species, while D. longicaudata originates in the Indo-Pacific region (Clausen 1978), Therefore, they would not have evolved niche divergence to avoid competition. D. areolatus was established in large numbers after being introduced to southern Florida (Baranowski and Swanson 1970), but has subsequently diminished and possibly disappeared from the region of its introduction. The current distribution pattern suggests that conqjetition by D. longicaudata may have caused its displacement. Note that D. longicaudata was introduced three years following the introduction of £>. areolatus. Thus D. areolatus had time to become estabUshed and migrate to fevorable locations to the north prior to the estabUshment of D. longicaudata. It is intriguing to think that had the sequence of introductions been reversed, D. areolatus may not have successfiUly been estabUshed. Similar cases involving fiaiit fly parasitoids in Hawaii are considered among the classic examples of apparent competitive displacement. Psyttalia (Opius) humilus (Silvestri) was the dominant parasitoid of the C. capitata in 1915, and was replaced by Diachasmimorpha tryoni (Cameron) from 1916 onward (Pemberton and Willard 1918). This displacement was apparently due to competitive superiority of the first-instar larvae of the latter species. During the late 1940s, several parasitoid species were released for the control of B. dorsalis (Clausen et al. 1965). Initially, D. longicaudata was the

PAGE 70

62 dominant species, only to be replaced by Fopius (Biosteres) vandenboschi (Fullaway), which was in turn replaced by Fopius (Biosteres) arisanus (Sonan) (van den Bosch et al. 1951). Two contributing mechanisms were offered for this displacement. First, the displacing species attack progressively earUer immature stages, which would be more prone to parasitism because of their proximity to the fruit surfece. Second, F. arisanus larvae appeared to inhibit the development of the other species, while F. vandenboschi larvae also inhibit development of D. longicaudata (van den Bosch and Haramoto 1953). In contrast to the report of van den Bosch and Haramoto (1953), several studies indicate that D. longicaudata may have a competitive advantage over other parasitoid species in situations of multiparasitism. Palacio et al. (1991) found that D. longicaudata was a superior competitor to both F. arisanus and Fopius (Biosteres) persulcatus Silvestri, indicating physical competition among first-instar larvae. Ramadan et al. (1984) suggested a similar advantage of D. longicaudata over D. tryoni in C. capitata hosts. Studies by Bautista and Harris (1997) with D. longicaudata and Psyttalia incisi (Silvestri) indicate that the sequence of oviposition is important, with the first parasitoid species to which the host is exposed having an advantage. However, while exposure first to P. incisi resulted in 77% of the progeny bemg of this species, the reverse sequence resulted in 99% of the progeny being D. longicaudata. These studies support the possibility that D. longicaudata may be a superior competitor to D. areolatus in multiparasitized hosts. To summarize, the two major phenomena observed in this study are the absence of D. longicaudata in the interior region of central Florida, and the absence of D. areolatus in portions of southern Florida. Which fectors may affect the interactions

PAGE 71

63 between D. areolatus and D. longicaudata, and how could this result in the observed pattern of distribution? Sivinski et al. (1998) hypothesized that the co-occurrence of both species at LaBelle may be the result of "counter-balanced competition" (cf Zw6lfer 1971) where D. areolatus is superior to D. longicaudata in locating host patches (=extrinsic competitor) and D. longicaudata is superior in exploiting these patches (=intrinsic conqjetitor). The better searcher would be at an advantage at locations that have a less predictable supply of hosts in time or space, while the better intrinsic conq)etitor would benefit from more homogeneous host availability. In the more northern interior regions of Florida, where temperatures are more variable, large gaps may occur between fruiting cycles of the various hosts, and in particular between the fall fruiting of guava and the spring fruiting of loquat. At coastal locations where temperature conditions are more homogeneous, trees may have more than one fruiting cycle, filling in the temporal gaps in fruit availability (Nguyen et al. 1992). Furthermore, additional tropical host fruits occur in the southern coastal regions (see Hennessey 1994). The former conditions would fevor the superior searcher, presumably D. areolatus, while the latter would benefit the superior intrinsic competitor, i.e., D. longicaudata. In extreme conditions one parasitoid species may driven to extinction, and at intermediate locations both would persist. At LaBelle large temporal gaps in hosts may be balanced by spatial abundance, in particular of guava, enabling the persistence of a sizable population of D. longicaudata. Diapause development is an important mechanism allowing insects to cope with periods of low host availability. There is evidence that D. longicaudata individuals do indeed enter diapause (Aluja et al. submitted, Ashley et aL 1976, Clausen et al. 1965,

PAGE 72

64 Chapter 5). However, in Mexican populations both the proportion of individuals entering diapause and the length of diapause period are greater for D. areolatus than for D. longicaudata (Aluja et al. submitted). Additionally, Aluja et al. (submitted) report circumstantial evidence that D. areolatus adults may enter a reproductive diapause. These observations suggest that D. areolatus populations may be able to better survive long periods without hosts. The presumed competitive advantage of D. longicaudata over D. areolatus may be the result of other mechanisms, besides larval con^etitioa Diachasmimorpha longicaudata may have an advantage in locating fruits containing host larvae, or in locating larvae within fruits. Diachasmimorpha longicaudata locates hosts within fruits by sensing the vibrations of the feeding larvae (Lawrence 1981). In the laboratory D. longicaudata females can locate hosts without the presence of host fruit odors. On the other hand, host larvae alone are not attractive to D. areolatus females, and addition of fresh fruit odors is sufficient to stimulate oviposition behavior in this species (Chapter 7). It is possible that D. areolatus may respond to host vibrations following exposure to host fruit odors. However, its dependence on fruit odors suggests that vibrations may be a relatively less important stimulus for D. areolatus than for D. longicaudata, and consequently it may be at a disadvantage in locating larvae within fruit. Conversely, this dependence on fruit odors may indicate a superior ability of D. areolatus to locate host patches. Note, however, that D. longicaudata females are attracted to volatiles associated with rotting fruit (Greany et al. 1 997). Additionally, D. longicaudata would have a conq)etitive advantage if it were more fecund. In the laboratory, D. longicaudata can produce a large number of progeny

PAGE 73

65 over a short period of time, with most eggs laid within a few days (Greany et al. 1976). On the other hand, D. areolatus appears to produce smaller numbers of progeny over longer periods of time (Chapter 6). If this is the case in the field, an individual D. longicaudata female, after locating the host patch, could exploit it at a fester rate than a D. areolatus female. Finally, the ovipositor of D. longicaudata is longer than that of D. areolatus {D. longicaudata, 5.49 ± 0.21 mm, range 4.67-6.40 mm, n=7; D. areolatus, 3.80 ± 0.11 mm, range 3.43-4.11 mm, n=7; t=7.18, p<0.0001). The ovipositor of £>. longicaudata is also consistently longer in relationship to body size as estimated by the ratio between ovipositor length and wing length (t= 14.59, p<0.0001, n=7). This enables D. longicaudata to reach larvae deeper within fruits, thus allowing access to a larger proportion of larvae, especially in large fruits (Sivinski et al. 1997). As discussed above, D. longicaudata distribution may be Umited to the north by (1) direct effects of low winter temperatures or (2) periods of low host availability. There is some evidence in support of both hypotheses. The two effects could act in concert, and may not be mutually exclusive. Some support for the second hypothesis can be found in the literature. PreUminary observations in Mexico, suggest that D. longicaudata is less common at low altitudes (M. Aluja, pers. comm). This is the opposite of what would be expected if this species was adversely affected by low temperatures. The low altitude habitats are drier, and as a result there are larger gaps in host availability. Thus, the temporal availability of hosts at low altitudes in Mexico is similar to that in the colder regions of Florida.

PAGE 74

66 In Mexico, D. longicaudata is the dominant species in an area of mixed cukivation, while it is absent in native habitats where D. areolatus is most common (Aluja et al. 1990, Hernandez-Ortiz et al. 1994). Similarly, in Amazonas State, Brazil, D. areolatus is the dominant parasitoid in rural locations while Opius sp. nr. bellus is dominant in urban areas (Canal D. et al. 1995). Native habitat is more heterogeneous in host availability, favoring a superior searcher, while there is a more predictable supply of hosts in cultivated or urban areas, which would favor the better mtrinsic con5)etitor. In both cases, the presence of wild hosts may give D. areolatus refuge from competition, thus preventing displacement. Such a refuge does not exist in Florida, where most hosts are in either urban or agricultural habitats. Feral populations of guavas exist in some areas, but fruiting of these trees is mostly limited to late summer. The biology of U. anastrephae is little known, and its interactions with other parasitoid species unclear. Laboratory rearing data from Brazil suggest that it attacks the same larval instars as D. areolatus and D. longicaudata (R. Sugayama, pers. comm.). In Mexico, Sivinski et al. (1997) observed negative relationships between U. anastrephae and D. areolatus within tree canopies. This was interpreted as being the possible result of evolution of divergent niches in these sympatric species, which would reduce direct con^tition. However, the inverse among-site relationship observed between these two species in this study suggests that significant within-site competition may be occurring. Note that U. anastrephae is common only in small fruits such as Surinam cherry (Table 3-6). Thus significant competition would only occur in such fruits. As D. areolatus was established in Florida with U. anastrephae already present, in appears that competition by U. anastrephae alone is not highly significant. However, U. anastrephae occurs in large

PAGE 75

67 numbers in most of the same towns where D. longicaudata is common (Figures 3-4 through 3-6). Thus, coupled with the competitive pressure of D. longicaudata, it may have contributed to the displacement of D. areolatus. Alternatively, U. anastrephae may be common at these locations because D. longicaudata had suppressed D. areolatus, thus releasing the former species from competition. There has been some debate over the merits of multiple introductions of natural enemies in biological control programs. Some workers have suggested that interspecific competition may reduce the overall level of host suppression, while others claim that additional species would increase control levels (see discussion in Van Driesche and Bellows 1996). DeBach and Rosen (1991) state that "displacement of a fairly effective estabUshed natural enemy species by another imported species means that the second one is more effective, and will produce even better host population regulation". Could the displacement of D. areolatus by D. longicaudata have reduced the suppression of the Caribbean fruit fly? This scenario appears to be possible. Mean parasitism levels for D. areolatus in Surinam cherry surpassed 20% at four towns, and at LaBelle reached 36% (Table 3-9). Meanwhile, mean parasitism for D. longicaudata was no higher than 15% (at Miami). Even with D. longicaudata and U. anastrephae combined, the highest mean parasitism level was 20% (at Belle Glade). Thus it appears that D. areolatus may be able to contribute to higher levels of control than the other parasitoid species, possibly due to its greater searching efiSciency. If this is the case, how then could this more efficient parasitoid be displaced? We could attempt to understand parasitoid population dynamics on a regional scale by examining the possible dynamics within individual host trees. Consider two Surinam

PAGE 76

68 cherry trees. The &st produces large numbers of fruits over a short period of time. Host larvae would be abundant, and parasitoids would be limited by the nimiber of fruits and larvae that they could locate. As more individuals of the superior searcher would be able to locate the resource quickly, on a population level it would be capable of locating greater numbers of larvae. The second tree produces fruits over a long period of time in limited patches. The superior searcher would have the initial advantage of colonizing the resource more quickly. However, after the superior intrinsic con^titor finally locates the resource, the latter species would have the advantage. While over time the former species may parasitize on average a higher proportion of hosts, towards the end of the fruiting cycle the latter species would dominate and possibly displace the superior searcher. Such dynamics (though not leading to displacement) where observed within trees at LaBelle, with both the relative and absolute numbers of D. longicaudata increasing over time (Sivinski et al. 1998). Note that parasitism by D. longicaudata can approach 100% at the end of the fruiting period in fruits such as Surinam cherry. If the majority of host trees within a town were of the second type, total displacement could occur. Fruiting patterns within trees are dependent on weather conditions, and could change over time. It is conceivable that in certain years most trees would be of the second type, and the population dynamics would lead to displacement. In other years the first type may dominate, and if displacement had previously occurred, the remaining parasitoid would not sufficiently respond to the growing host population, and parasitism levels would be low. In conclusion, a superior intrinsic competitor may displace a superior searcher, leading to a reduction in host suppression.

PAGE 77

69 D. areolatus is absent or rare at northern coastal locations, even though total parasitism at these sites is low. It appears that other fectors besides competition must account for this absence. This suggests that unidentified environmental factors associated with coastal locations may have contributed to its disappearance from southern Florida as well. However, if coastal conditions were unsuitable to D. areolatus, it would not have been ejqiected to become established in large enough nimibers to enable it to spread to distant regions of the state. Perhaps widespread pesticide applications against mosquitoes and other biting insects, which is most prevalent in coastal regions, contributed to its disappearance. Could D. areolatus have displaced D. longicaudata in central Florida, just as D. longicaudata may have displaced D. areolatus in the south? D. areolatus obviously migrated from its original release area in southern Florida to the areas in which it currently dominates in the central part of the state. However, it is unclear whether it was present there when D. longicaudata was released in 1972, only three years after its own introduction. Could a superior searcher displace a superior intrinsic competitor? Although a superior searcher would have a relative advantage in a situation of less predictable hosts, allowing it to be more successM on a population level, the superior competitor by definition would have an advantage when both species occur together on a patch, regardless of the mechanism of competition. It is more likely that the superior con5)etitor would be driven to extinction due to lack of hosts than because of effects of a less competitive parasitoid species. Is it advisable to release D. longicaudata in periodic inundative releases in the regions of central Florida from which it is presently absent? It is quite easy to rear D.

PAGE 78

70 longicaudata in the laboratory, and mass-rearing can be achieved at relatively low cost. In contrast, D. areolatus is much more difficult to rear (Chapter 6). Therefore, it would be cost-effective if the program of inundative releases of D. longicaudata could be expanded to all regions of Florida (see Bums et al. 1996). There is evidence that largescale inundative releases of D. longicaudata could reduce host populations (Sivinski et al. 1996). The numbers of parasitoids released is presumably much higher than those naturally occurring in the field (Kniplmg 1992). Thus, there is no real difference between augmenting existing populations and releases in areas where parasitoids do not occur. There is no reason to beUeve the parasitoids would not be as effective in all regions of the state, at least during warm periods of the year. However, there is a substantial risk that these releases would cause the permanent displacement of D. areolatus. If releases are terminated after such a displacement, D. longicaudata would not be expected to become permanently established (because they would not survive the winter or periods lacking in hosts), and no parasitoids would remain. Thus mitiation and subsequent termination of an inundative release program for D. longicaudata could ultimately lead to an explosion of A. suspensa populations. If inundative parasitoid release in central Florida is pursued, it may be more advisable to develop more cost-effective rearing procedures for D. areolatus, with the objective of releasing this species in areas where it currently occurs.

PAGE 80

CHAPTER 4 LOCAL TEMPORAL AND SPATIAL DISTRIBUTION PATTERNS OF DIACHASMIMORPHA LONGICAUDATA AND DORYCTOBRACON AREOLATUSm AN AREA OF CO-OCCURRENCE The distributions of Diachasmimorpha longicaudata and Doryctobracon areolatus overlap in Florida only within a limited region (Chapter 3). The town of LaBelle, situated between Lake Okeechobee and the Gulf of Mexico, is one of the few locations in which both are common (Chapter 3, Sivinski et al. 1996, 1998). Studies of the temporal and spatial dynamics of these parasitoids at LaBelle may help explain how they co-occur at this location, while in most areas of Florida they do not. More specifically, they could address hypotheses generated in Chapter 3, i.e. that low temperatures have a direct or indirect negative effect on D. longicaudata, and that interspecific competition could partially explain the disappearance of D. areolatus from southeastern Florida. Tenqx)ral and spatial dynamics of these species within trees were studied by Sivinski et aL (1998, pers. comm.). This chapter examines similar dynamics on a larger scale, by comparing parasitoid abundance among trees within LaBelle. Additionally, I examined the temporal dynamics of the population as a whole, both within and among years. 72

PAGE 81

73 Materials and Methods Fruit Sampling in 1996 Loquat (Eriobotrya japonica (Thunb.)) and Surinam cherry {Eugenia uniflora L.) fruits were sampled at LaBelle and the adjacent town of Ft. DeNaud every two weeks from Week 4 (late January) to Week 22 (early June) of 1996. Most loquats were sampled from Week 4 to Week 14, but some were available until Week 18. Most Surinam cherries were sanq)led from Week 14 to Week 22, with one sample each collected m Weeks 4 and 6. Every tree within the towns which was found to have at least ten fruits was sanpled. Each sample inchided fruits from a single tree, A total of 256 san^les was collected, ranging from 3 in Week 4 to 51 in Week 16. Numbers of fruits per sample ranged between 17-106 for loquat, and between 18-151 for Surinam cherry. These numbers represented either all fruits present on the tree, or the maximal number that could be put in a single layer on the screen within the bucket (see Chapter 3 for details of fruit handling after collection). Fruits were sampled randomly from different parts of the tree. Each fruit sample was weighed following collection Upon intensive collection of Surinam cherry fruits, it became apparent that they were infested by large numbers of fungal spores. Therefore, beginning in Week 18, fruits were washed in a 0.03% sohition of sodium benzoate. Abundance of the three parasitoid species, D. areolatus, D. longicaudata and Utetes anastrephae, was compared between LaBelle and Ft. DeNaud. Due to significant differences between the two towns (see Table 4-1), aU subsequent analyses included only

PAGE 82

74 data from LaBelle. Additionally, because U. anastrephae was relatively uncommon at LaBelle, in was not considered in these analyses. Analysis of Distribution Among Trees Spatial distributions of Caribbean fruit flies, Anastrepha suspensa (Loew), and their parasitoids among trees were examined using data collected in LaBelle during 1996. Distribution was visualized using the Surfer software program (Golden Software) and the kriging method. Kriging uses sampled data to produce a grid of estimated values quantifying the entire distribution of the parameter of interest. Ultimately, kriged data are used to create iso lines of equal parameter density visualized as a 2-dimensional contour map. For a detailed description of spatial analysis and its use in entomology see Brenner et al. (1998). Lx)ngitude and latitude coordinates for each host tree were obtained using the Microsoft Automap Street Plus software program, and adjusted sUghtly to fit a TIGER/Line base map (U.S. Census Bureau). Figure 4-1 illustrates the quadrants which are included in this analysis. The town of LaBelle, situated to the south of the Caloosahatchee River, can be divided into four quadrants by State Road 80 transecting from east to west, and by State Road 29 running from north to south. A fifth section is in the town of North LaBelle, to the north of the river. Several samples collected in other quadrants were not included in this analysis. Because of the low occurrence of parasitoids in loquat, only Surinam cherry samples were considered in the spatial analysis. Separate maps were produced for each sampUng period, from Week 14 to Week 22. Relationships within each samphng period between fly abundance and that of each parasitoid species, and between parasitoid species, were examined using regression or correlation analysis.

PAGE 83

NolliLaBdle < > 1000 meters Figure 4-1. Map of LaBelle, Florida, showing quadrants sampled in study.

PAGE 84

76 Temperature Measurements Parasitoid abundance may be influenced by local variability in ten^)eratures. In order to assess the occurrence of spatial variability in winter ten5)eratures, five Optic StowAway data loggers (Onset Con:?)uter Corp.) were placed fi-om December 1996 through March 1997, three at LaBelle and two at Ft. DeNaud. Loggers were placed one meter above the ground on the northern side of major limbs of large Surinam cherry trees, from which a significant number of parasitoids had been recovered during the previous spring. Loggers were set to record the temperature every 24 minutes. Mean, mean minimum and mean maximimi temperatures were calculated for each month. These variables were subjected to Friedman's two-way analysis for block designs. This was achieved with the SAS software program by obtaining ranking among sites within days and then performing an analysis of variance on these ranks among sites (SAS Institute 1982). Means were subsequently compared with the Waller-Duncan k-ratio t test. In addition to the above mentioned variables, extreme minimum and maximum temperatures were noted. Comparisons of Parasitoid Abundance Among Years Various studies on parasitoids of A. suspensa have been conducted at LaBelle during recent years. In addition to data from the current study from 1995 (see Chapter 3) and 1996, Sivinski et al. (1996, 1998) sanq)led parasitoids for various purposes during each of the years 1991 through 1994. Thus, comparisons of parasitoid abundance among years could be made, and relationships with environmental factors examined.

PAGE 85

--V . 77 In 1991 and 1992, samples were collected in a feshion similar to the present study; No more than one sample was collected from each tree in a single week. No manipulations (besides transformation) were performed on these data prior to analysis. In 1994, multiple samples were collected from each tree in a single day, and often trees were sampled more than once a week. In order to make reUable con^)arisons with other years, these data were manipulated in the following manner. Samples collected from a single tree on a single day were combined by summing the numbers of parasitoids and flies emerging. Parasitism levels were calculated for each tree each day. Where trees were sampled more than once a week, mean parasitism per tree per week was calculated, and was considered to be a single sample for analysis. In 1993, two separate studies were performed, one as in 1991-1992 (single sample per tree per week) and the other as in 1994 (multiple samples per tree). Where multiple samples were collected, data were manipulated as with the 1994 data. The resulting parasitism levels per tree per week were considered single samples, and given equal weight in the analysis as samples from the other study. Note that, as mentioned above, only samples from LaBelle were considered. Thus, the 1995 data used in this chapter is a subset of the "LaBelle" data given in Chapter 3, which includes Ft. DeNaud. Associations of parasitoid abundance with environmental fectors were examined by linear regression analysis. Temperature and precipitation data were obtained from the Southeast Regional Chmate Center, Columbia, South CaroUna. Parasitism by each parasitoid species in loquat or Surinam cherry was related with the foUowing variables: mean temperature, mean and extreme minimum temperatures, mean and extreme

PAGE 86

78 maximum temperatures, and precipitation. Separate analyses were performed for conditions prevalent during each month preceding fruit collection Thus, parasitism in loquat was related with tenq)erature and precipitation variables for the months October through February, and parasitism in Surinam cherry was related with these variables for the months October through April. Results Comparison of Abundance Between LaBelle and Ft. DeNaud The various parasitoid species varied greatly m their abundance between LaBelle and Ft. DeNaud during 1996 (Table 4-1). Parasitism by D. areolatus in loquat averaged 3% at LaBelle, but it was not collected at Ft. DeNaud. More D. longicatuiata were recovered from loquat at Ft. DeNaud than at LaBelle, but the difference was not significant. U. anastrephae was uncommon in loquat at both towns. The differences between the towns were pronounced in Surinam cherry; D. areolatus was more common at LaBelle than at Ft. DeNaud, while both D. longicaudata and U. anastrephae were more abundant at Ft. DeNaud than at LaBelle. Note that these towns border each other, and the eastern most sample included in this study from Ft. DeNaud is only 2.2 km away from the western most sample from LaBelle. Because of these differences between the towns, only data from LaBelle were included in subsequent analyses. .

PAGE 87

79 Table 4-1 . Comparisons of percent parasitism by various parasitoid species in 1996 between the towns of LaBelle and Ft. DeNaud. Hnitf' rruii LaBelle Ft. DeNaud t P species n Mean (SE) n Mean (SE) 50 2.87 (1.02) 29 0 '5 CO 3.58 0.001 D lon^cQudcitn 50 0.09 (0.07) 29 0.61 (0.37) U. 10 IJ. anastrephae 50 0.02 (0.02) 29 0.03 (0.03) -0.30 0.77 Surinam D. areolatus 117 30.15(3.02) 20 5.87 (3.51) 5.12 0.001 cherry D. longicaudata 117 1.86 (0.49) 20 11.18(3.57) -3.02 0.007 U. anastrephae 117 0.21 (0.09) 20 3.66(1.52) -2.41 0.026 Distribution Among Trees Densities of A. suspensa were generally low at the beginning of the Surinam cherry season. During Week 14, flies were concentrated at two focal points, one in the northeastern quadrant of LaBelle, and the second in North LaBelle, just north of the Caloosahatchee River (Figure 4-2). During Week 16, fly densities were again high at these locations. However, high infestations were observed also just to the south of the river, in the southern region of the northwestern quadrant, and at tiie southern fringes of town (Figure 4-3). Infestation levels were generally higher during Week 18, with highest numbers observed in North LaBelle, the northeastern quadrant just south of flie river, and the northwestern quadrant (Figure 4-4). During Week 20, fly numbers were high at most locations, with highest infestations observed in the southwestern quadrant (Figure 4-5). A similar pattern was observed during Week 22, with focal points in the southwestern and northwestern quadrants (Figure 4-6).

PAGE 88

80 D. areolatus was uncommon at the beginning of the Surinam cherry seasoa During Week 14, only small numbers were recovered from two host trees (Figure 4-7). Note that these were the same trees which had the highest infestations of A. suspensa (Figure 4-2). By Week 16, D. areolatus was recovered from 11 of 35 host trees. Four focal points were observed, in the northwestern, northeastern and southwestern quadrants (Figure 4-8). Parasitism levels increased dramatically by Week 18, with highest levels (over 50%) observed in the general vicinity of the focal pomts in the previous sampling period (Figure 4-9). Parasitism of over 50% was widespread during week 20, reaching over 80% in three areas: North LaBelle, the northeastern and extreme northwestern quadrants, and the southwestern quadrant (Figure 4-10). A similar pattern of generally high parasitism levels was observed during Week 22 (Figure 4-11). ^ i D. longicaudata was not recovered from Surinam cherry at LaBelle until Week 16, when it was found in a single host tree just south of the river (Figure 4-12). During Week 18 it was recovered from the same tree, and one additional tree in the southwestern quadrant (Figure 4-13). Parasitism increased dramatically by Week 20, when it was recovered from 1 1 of 28 hosts. Four focal points with over 10% parasitism were apparent, two each in the northwestern and southwestern quadrants (Figure 4-14). During Week 22, highest parasitism levels were observed in the southwestern quadrant, with a second area of parasitism apparent on both sides of the Caloosahatchee River (Figure 4-15). Parasitism by D. areolatus was significantly related with A. suspensa infestation levels during Week 14 (R^=0.68, F=10.86, p=0.02) and Week 16 (R^=0.45, F=23.67, p<0.0001), but not Week 18 (R'=0.11, F=2.73, p=0.11). Week 20 (R'=0.07, F=1.95, p=0.17), or Week 22 (R'=0.02, F=0.19, p=0.67). Parasitism by D. longicaudata vfas not

PAGE 89

81 significantly related with A. suspensa infestation levels during any week (Week 20, RM.03, F=0.91, p=0.35; Week 22, R^= 0.10, F=1.09, p=0.32). The ratio between parasitism by D. longicaudata and that by D. areolatus mirrors the parasitism by D. longicaudata alone during Weeks 20 and 22 (Figures 4-16 and 4-17, conspdXQ with Figures 4-14 and 4-15). This is probably due to the relatively even distribution of D. areolatus during these weeks (Figures 4-10 and 4-11). Parasitism by the two species was significantly correlated during Week 18 (R=0.44, p<0.03). With all samples considered, there were no significant relationships between parasitism levels of D. areolatus and D. longicaudata during any other week (Week 16, R=0.10, p=0.60; Week 20, R=0.03, p=0.87; Week 22, R=0.28, p=0.35). However, when considering only samples from which D. longicaudata was recovered, there was a significant negative relationship during Week 20 (Figure 4-18), Temperature Measurements Considering the apparent negative relationship between cold winter temperatures and presence of this species (Chapter 3), it was hypothesized that the abundance of D. longicaudata at Ft. DeNaud relative to LaBelle may be the result of warmer winter temperatures. It was furthermore hypothesized that the river may have a moderating effect on temperatures. Temperature variables obtained at five sites at LaBelle and Ft. DeNaud are detailed in Table 4-2. The three coldest locations in all months, in terms of extreme minimum temperature, were the two locations at Ft. DeNaud and the site at LaBelle closest to the river. In terms of mean minimum temperatures, both Ft. DeNaud sites were significantly colder than two of the three LaBelle sites (Table 4-3). Contrary to expectations, the LaBelle site ferthest from the river was the warmest location in terms of

PAGE 90

82 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Pupae / gram fruit Figure 4-2. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 14th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 7 of 11 hosts.

PAGE 91

83 -81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Pupae / aram fmit Figure 4-3. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 16th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 31 of 35 hosts.

PAGE 92

84 " I — i ' — T 1 -81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Pupae / gram fruit Figure 4-4. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 18th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 24 of 25 hosts.

PAGE 93

85 — r — "—^ 1 \ 1 — -61.445 -81.440 -81.435 -81.430 SI.AZS Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Pupae / gram fmit Figure 4-5. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 20th week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 28 of 28 hosts.

PAGE 94

86 -81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Pupae / gram fruit Figure 4-6. Spatial distribution of Caribbean fruit fly infestation of Surinam cherry fruits at LaBelle during the 22nd week of 1996. Circles indicate locations of hosts sampled. Flies were recovered from 14 of 14 hosts.

PAGE 95

87 -81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Parasitism level Figure 4-7. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 14th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 2 of 11 hosts. Parasitism level is the ratio between the number of Z). areolatus emerging and the number of all parasitoids and flies emerging.

PAGE 96

88 1 1 1 " 1 J — -81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Parasitism level Figure 4-8. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 16th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 1 1 of 35 hosts. Parasitism level is the ratio between the number of D. areolatus emerging and the number of all parasitoids and flies emerging.

PAGE 97

-81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Parasitism level 1.0 Figure 4-9. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 18th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 15 of 25 hosts. Parasitism level is the ratio between the number ofZ). areolatus emerging and the number of all parasitoids and flies emerging.

PAGE 98

26.755 ., V . . — _J , 1 1 r -81.445 -81.440 -81.435 -81.430 Longitude (degrees) -81 425 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Parasitism level Figure 4-10. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fhiits at LaBelle during the 20th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 27 of 28 hosts. Parasitism level is the ratio between the number oiD. areolatus emerging and the number of all parasitoids and flies emerging.

PAGE 99

91 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Parasitism level Figure 4-11. Spatial distribution of parasitism by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 22nd week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 13 of 14 hosts. Parasitism level is the ratio between the number of£>. areolatus emerging and the number of all parasitoids and flies emerging.

PAGE 100

92 26.775 26.770 in c» 26.765 I 26.76026.755 -81.445 -81.440 -81.435 -81.430 Longitude (degrees) -81.425 0.00 0.01 0,02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 Parasitism level Figure 4-12. Spatial distribution of parasitism by Diachasmimorpha longiccmdata in Surinam cherry fruits at LaBelle during the 16th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 1 of 35 hosts. Parasitism level is the ratio between the number of D. longicaudata emerging and the number of all parasitoids and flies emerging.

PAGE 101

Figure 4-13. Spatial distribution of parasitism by Diachasmimorpha longicaudata in Surinam cherry fruits at LaBelle during the 18th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 2 of 25 hosts. Parasitism level is the ratio between the number of D. longicaudata emerging and the number of all parasitoids and flies emerging.

PAGE 102

94 26.775-81.445 -81.440 -81.435 -81.430 Longitude (degrees) -81.425 (^y^O.O^O^O^O o o o o o o o o o o Parasitism level Figure 4-14. Spatial distribution of parasitism by Diachasmimorpha longiccmdata in Surinam cherry fruits at LaBelle during the 20th week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 1 1 of 28 hosts. Parasitism level is the ratio between the number of D. longicaudata emerging and the number of all parasitoids and flies emerging.

PAGE 103

95 ^^<>j^O^O^O^O o o o o o o o o o o Parasitism level Figure 4-15. Spatial distribution of parasitism by Diachasmimorpha longicaudata in Surinam cherry fhiits at LaBelle during the 22nd week of 1996. Circles indicate locations of hosts sampled. Parasitoids were recovered from 8 of 14 hosts. Parasitism level is the ratio between the number of D. longicaudata emerging and the number of all parasitoids and flies emerging.

PAGE 104

96 I r^— ^ 1 -81.445 -81.440 -81.435 -81.430 -81.425 Longitude (degrees) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 D. longicaudata / D. aneolatus Figure 4-16. Contour map of the ratio between parasitism by Diachasmimorpha longicaudata and that by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 20th week of 1996. Circles indicate locations of hosts sampled.

PAGE 105

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 D. longicaudata / D. areolatus Figure 4-17. Contour map of the ratio between parasitism by Diachasmimorpha longicaudata and that by Doryctobracon areolatus in Surinam cherry fruits at LaBelle during the 22nd week of 1996. Circles indicate locations of hosts sampled.

PAGE 106

98

PAGE 107

99 00 CM 00 NO fN CM s O 00 * tN !f-' t-~' o w-i E eg c o Q S e2 00 >/~i so o oo ^ On COn Tj^ ro On ^• 00 Tiro ro CO CM — < CM ^. fo-1 CM <— I ro 0\ CM po >r> so ? S ON OCMi— it^cMNOrorot-so u arro OS Os ON 00 o 00 p <4CO CM NO O NO o ^' x> E a> u Q o Z E III JO E V u o Q 12 1 a> 3 ce 1

PAGE 108

100 2 I to I s 3 x n 2 I s 3 E .S 3 IS I 2 C/1 x: o U < OQ U o o r(« o 1—1 00 00 CO Q < OQ 5 o ON x> V 00 Uh (N ary Q U OQ 3 O 1 00 OO oo OS >/-) s Q U <: CJ OQ o O >o 00 00 Q (N ro CQ U < u U oo 00 00 so 2 09 (J < u t o (0 b I-H — 1 m r-i e OQ <(! (J (J o 00 0^ ^1 Q (N (N t Q u < OQ CJ v s> U S OQ u < OQ <1> J3 ^ o o ^ Id O O _o _o C3 "« U U o o s ^ o O o o C (/) S S O ^ r~ "53 "aj OQ OQ ^ J3 u,« > ''-I O > 2^ ID V O O o 2 I -I o « lU S •T3 ^ 3 OQ " CO 4J u o _ §^ -I s o a> a> (u •tn ."tt C/3 C/J «/l C« (1> 3 2 o c § ^ g o o -rf -0' ^ 2 2 o II ^ " o w -S in 0) ^ ^ w s

PAGE 109

101 mean, mean minimum and mean maximum temperatures for all months. Interestingly, the two Ft. DeNaud sites were warmer than the other two LaBelle locations for all months in terms of mean maximum temperature. With the exception of a possible positive effect of high maximum temperature on D. longiccnidata, these measurements failed to explain the differences in parasitoid abundance between the two towns. Note, however, that these measurements were performed the winter following fruit sampling in this study. Thus, it is possible that they are not representative of the true situation during the previous year. For the LaBelle sites, there is a possible relationship between minimum temperatures and parasitism by D. longicaudata. At the warmest site (81.444 West, 26.759 North, 1000 meters south of river) parasitism by D. longicaudata was highest (26.3% during Week 20). At the coldest site (81.436 West, 26.762 North, 750 meters south of river) D. longicaudata was not recovered. The third site (81.441 West, 26.771 North, 100 meters north of river) was intermediate in terms of both temperature and parasitism (4.7% during week 22). Thus, a local effect of low winter temperatures cannot be dismissed. However, temperatures need to be measured at a larger number of sites in order to substantiate such a relationship. Comparisons Among Years The pattern of parasitism by D. areolatus was somewhat similar for all years (Figure 4-19). Mean parasitism reached a peak of 30-60% in May or June, corresponding to the peak of Surinam cherry availability. This was expected, as parasitism rates are usually higher in Surinam cherry than in loquat or guava (see Table 3-6). Mean parasitism was higher in loquat (early spring) during 1991-1993 than during 1994-1996. In all three years there was significantly higher parasitism than in 1996. However, only in

PAGE 110

102 1993 was it significantly higher than in 1994 or 1995 (Figure 4-21). There was an apparent gradual decrease in parasitism in Surinam cherry from 1991 through 1994, followed by an increase in 1995 and decrease again in 1996. However, these differences among years were not significant (Figure 4-22). Temporal patterns of parasitism by D. longicaudata were much more variable (Figure 4-20). Peak abundance in spring for various years ranged from 4% in 1996 to 23% in 1994, and was observed in either April, May or June. Note the very similar patterns in 1991 and 1996. In both years, parasitism in early spring was at or close to zero, with mean parasitism reaching a peak of less than 5% in June. The unusually high peak in August 1991 should be treated with caution, as it is a mean of only four samples. Parasitism in loquat was significantly higher during 1995 than during either 1991, 1994 or 1996 (Figure 4-21). Parasitism in loquat during 1992 and 1993 was not significantly different than in any other year. In Surinam cherry, parasitism during 1993, 1994, and 1995 was significantly higher than during other years, while parasitism in 1995 was higher than in all other years except for 1994 (Figure 4-22). Mean parasitism by D. areolatus in loquat was higher than that by D. longicaudata for all years except 1995 (Figure 4-21), but the difference was significant only during 1991 (paired-comparisons t test, t=2.86, p<0.02), 1993 (t=5.23, p<0.0001), and 1996 (t=3.55, p<0.0009), and not 1992 (t=1.58, p=0.13), 1994 (t=2.24, p=0.09), or 1995 (t=1.37, p=0.18). Parasitism by D. areolatus in Surinam cherry was significantly higher than that by D. longicaudata for all years (1991, t=7.95, p<0.0001; 1992, t=8.59, p<0.0001; 1993, t=3.83, p<0.0005; 1994, t=4.68, p<0.0003; 1995, t=2.97, p<0.007; 1996, t=10.86, p<0.0001).

PAGE 111

103 70 T 1 60 January February March April May June July August September 70 TJanuary Febmary March April May June July August September Month Figure 4-19. Parasitism by Doryctobracon areolatus at LaBelle during the years 19911996

PAGE 112

104 January February March April May June July August September 40 T — . 35 January February March April May June July August September Month Figure 4-20. Parasitism by Diachasmimorpha longicaudata at LaBelle during the 1991-1996.

PAGE 113

105 Mean parasitism by D. longicaudata in loquat was significantly related Avith extreme minimum temperature of the previous December (Figure 4-23). Note that during the three years following a cold December, parasitism was at or close to zero. Mean parasitism by D. areolatus in loquat was significantly related with several environmental variables. The most significant was extreme minimum temperature of the previous January (Figure 4-24). Other significant variables included mean minimum January temperature (R^=0.68, F=8.60, p<0.05), mean January temperature (R^=0.75, f=11.73, p<0.03), and October precipitation (negative relationship, R^=0.86, F=25.17, p<0.008). There were highly significant relationships between parasitism by D. longicaudata in Surinam cherry and both mean and mean minimum December temperatures (Figures 4-25 and 4-26). Other variables significantly related with D. longicaudata abundance in Surinam cherry included mean November temperatures (R^=0.66, F-7.91, p<0.05), mean March temperature (R^=0.86, F=24.16, p<0.008), mean maximum November temperatures (R^=0.67, F=8.44, p<0.05), and December precipitation (R^=0.80, F=15.52, p<0.02). Mean parasitism by D. areolatus in Surinam cherry was significantly related with mean February temperatures (R^=0.76, F=12.77, p<0.03). The lack of highly significant relationships was not surprising, given the relatively high parasitism levels during all years, and the lack of significant differences among years.

PAGE 114

' 106 E Io> (D 0) CO CD < 1 CD o> I? §"2 2 II H .2 c3 -t-j CO «J i "1 0) V 0 -o .9 S c -5 1 ^ ? § a c ^ ^ S3 f~ tS !> o " I s . -5 '% '5 On C 00 ON 00 S 7 CXOn O u op S II I -o ^ I

PAGE 115

107 Lusjijsejed tusojad ue9|/\| 3^ CS It o S 0} I T3 C at !>5 5 o Is •T3 C £0 to ss I i <5 ON — c OS o CO On 3 c .00 c
PAGE 117

109

PAGE 120

112 Discussion Results of the large-scale geographical study (Chapter 3) suggest that cold winter temperatures may have a negative effect on D. longicaudata populations. Analyses of the variability in parasitism among years suggest that cold temperatures have some effect on both D. longicaudata and D. areolatus. While minimum January temperatures appear to influence parasitism by D. areolatus in loquat, populations of this species appear to rebound to the point that no significant differences are observed among years in parasitism in Surinam cherry. The effects of cold winter temperatures appear to have a more profound affect on D. longicaudata. In years with cold minimum December temperatures, parasitism by this species in loquat is at or near zero. In the coldest years, populations fail to recover fully, with mean parasitism in Surinam cherry staying below 5%. The apparent relationships between January temperatures and D. areolatus abundance on one hand, and between December temperatures and D. longicaudata abundance on the other, suggest that the two species may be influenced by different environmental factors, whether direct temperature affects or indirect influence of temperature through host availability. This study does not fully resolve which of these factors may be important. Only during 1995 was D. longicaudata more abundant (though not significantly) than D. areolatus in loquat. The 1994 fall season was unique among the years of this study in that an unusual crop of loquat occurred during this time (Tim Holler, pers. comm.). Thus, there was a supply of hosts bridging the usual gap between common guava in late summer and loquat in early spring, making the temporal availability of hosts

PAGE 121

113 more similar to that normally observed on the southeastern coast of Florida. These observations support the hypothesis suggested in Chapter 3 that D. longicaudata benefits from a relatively constant supply of hosts. An examination of distribution among Surinam cherry trees during 1996 may shed light on parasitoid population dynamics, at least for years following cold winters. D. areolatus was initially recovered only from hosts with highest A. suspensa densities. In subsequent weeks, however, there was no relationship between fly densities and parasitoid abundance. Given the increase in fly densities over time, this suggests that parasitoids are dependent on fly abundance only when densities are low. D. longicaudata was not recovered during the first collection week, and was found in a large number of trees only six weeks later, or four weeks after D. areolatus was widely distributed. These differential temporal dynamics of the two parasitoid species are similar to those reported by Sivinski et al. (1998) within individual trees. Both species appear to initially occur at several focal points, from which they disperse to other host trees. However, while D. areolatus is abundant in most trees at the peak of the fruiting period, D. longicaudata populations at this time are still concentrated around the focal points due to its late colonization. The significance of focal points in the dynamics of parasitoid populations, as well as effects of local temperatures at specific hosts, would depend on the degree of dispersal. Little is known about dispersal of fiiiit fly parasitoids. Messing et al. (1994, 1995) studied the short-range (10-40 m) dispersal of mass-reared parasitoids of three species, including D. longicaudata. Further studies on mediumand long-range dispersal in field populations is necessary. Note that abundance of the various parasitoid species differs

PAGE 122

114 greatly between LaBelle and Ft. DeNaud. This suggests that long-range (>1 km) dispersal may be limited. The relationship between D. areolatus and D. longicaudata appears to be quite complex. A positive relationship was observed during Week 18, i.e., the second week during which D. longicaudata was recovered. The two host trees from which D. longicaudata was collected also were among the highest in terms of parasitism by D. areolatus. This suggests that environmental factor(s) may be similarly influencing both presence ofZ). longicaudata and abundance ofD. areolatus during this week. However, such a relationship was not observed in subsequent weeks. During the peak of fruit availability (Week 20) parasitism levels of D. areolatus in samples containing D. longicaudata did not differ significantly from parasitism in samples from which D. longicaudata was absent (t=1.04, p=0.31). Thus, presence of Z). longicaudata in a particular tree appears to be unrelated with the abundance of D. areolatus. However, once D. longicaudata arrives at a host, it has an apparent negative effect on D. areolatus abundance. Hence the negative relationship between the two parasitoids when only trees with D. longicaudata present are considered (Figure 4-18). Lack of niche separation within tree canopies (Sivinski et al., in preparation) suggests a potential for competitive interactions. For a specific example, consider the Surinam cherry tree at coordinates 26.768 North, 81.439 West. During Week 18, it had one of the highest levels of parasitism by D. areolatus, at 61% (Figure 4-9). This was also one of the two hosts from which D. longicaudata was recovered that week (Figure 4-13). Parasitism by D. areolatus in this tree declined to 52% by Week 20, while increasing to over 70% in all adjacent hosts

PAGE 123

115 (Figure 4-10). At the same time, parasitism by D. longicaudata increased to 26%, while remaining uncommon in surrounding trees (Figure 4-14). Alternating warm and cold winters may contribute to the coexistence of the two parasitoid species at LaBelle. Warm winters, which may positively contribute to host availability, allow increases in D. longicaudata populations. If the frequency of warm winters was to increase, it would enable consistently high population levels, thus increasing the competitive pressure on D. areolatus. Under certain circumstances this could lead to the displacement of D. areolatus, as discussed in Chapter 3. Conversely, cold winters like that of 1995-1996 cause a severe reduction in populations of this species. If the frequency of cold winters was to increase, D. longicaudata populations may be suppressed to a level from which they could not recover, leading to extinction. These scenarios may explain the absence of D. areolatus in southeastern Florida, where winters are warmer than at LaBelle, and the absence of D. longicaudata in central Florida, where winters are colder, while they coexist in the intermediate climate of LaBelle.

PAGE 124

CHAPTER 5 EFFECTS OF TEMPERATURE ON IMMATURE DEVELOPMENT, ADULT LONGEVITY AND OVIPOSITIONAL ACTIVITY IN DIACHASMMORPHA LONGICAUDATA The temporal and spatial distribution patterns of Diachasmimorpha longicaudata (Ashmead) in Florida suggest that low temperatures may have a negative effect on this species (Chapters 3 and 4). One hypothesis proposed is that temperatures affect the parasitoids directly, causing increased mortality. The objective of this chapter is to determine the temperature tolerances of immature and adult D. longicaudata. Additionally, temperature effects on arrested immature development and adult ovipositional activity are examined. Materials and Methods Immature Development D. longicaudata used in this experiment were tenth generation of a stock descending from individuals collected at LaBelle, Florida. Fifty female parasitoids were placed in each of twelve 20 cm^ screen cages. An oviposition unit (see description in Chapter 6) containing several hundred late third-instar Caribbean fruit fly, Anastrepha suspensa (Loew), larvae was placed on top of each cage for ca. 3 hr. Upon completion of this exposure period, the larvae were placed upon moist fine vermiculite (15-20 ml water per 100 cm' vermiculite) in small plastic containers. Containers were covered with lids with fine-mesh posh fabric. Larvae were allowed to enter the vermiculite for pupation at 116

PAGE 125

117 25°C for several hours. Containers were subsequently transferred to Florida Reach-In® environmental chambers (GaflFney Engineering, Inc., Gainesville, Florida, see Walker et al. 1993). Containers were randomly placed in each of six chambers, set at constant temperatures of 11, 13, 15, 17, 19 or 21°C (±0.2°C) and 90% relative humidity. Additionally, a container with unexposed host larvae was placed in each chamber as a control. This procedure was repeated for a total of six days. On the first two days, oviposition units contained ca. 700 host larvae. Following a break of four days, parasitoids were exposed to hosts for an additional four consecutive days, with oviposition units containing ca. 300-400 host larvae. Two containers with larvae exposed to parasitoids were placed in each chamber on each of the six days, for a total of 12 samples, except 21°C where only 8 samples were placed in the chamber over the final four day period. Parasitoid emergence was recorded daily until ca. 19 weeks after the median exposure date, and then weekly until termination of the experiment. Containers were removed from most chambers and placed in a room at 25°C ca. 21 weeks after exposure. Those samples at 15 or 17°C remained for an additional 4 weeks. Weekly emergence counts continued until 29 weeks after exposure, when no more emergence was observed. Note that temperature control failed 13 weeks after exposure in the chamber set at 19°C. As a result the temperature dropped to approximately 12°C, and remained at this level until removal of the samples 8 weeks later. Following termination of the experiment, the vermiculite was sifted and the number of host puparia, with and without fijngi, was counted. Parasitoid emergence was

PAGE 126

118 estimated as the ratio between the number of parasitoids emerging and the number of puparia without fungi. Adult Longevity and Ovipositional Activity Insects were obtained from the mass-rearing facility, Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Gainesville, Florida. Parasitoids had been in colony for ca. 30 generations. Twenty newly emerged D. longicaudata, 10 females and 10 males, were placed in 20 cm screen cages, and provided within honey-agar and water. Eight of these cages were placed upon two shelves within each of six Florida Reach-In environmental chambers. The chambers were set at constant temperatures of 15, 19, 23, 27, 31, or 35°C (±0.2°C). All chambers were set at the same saturation deficit, corresponding to relative humidities of 44, 57, 66, 73, 79, and 83%, respectively. Parasitoids were presented with 100 irradiated late third instar /i. suspensa larvae for 70 min twice a week. Exposure was during the third hour of a 12 hr photophase, which preliminary experiments indicated was a period of relatively high ovipositional activity at room temperature. Mortality was recorded daily until the death of the last parasitoid. Ovipositional activity was estimated by counting the progeny resulting from the first two host exposures, at ages of 4 and 7 days. Progeny production is known to be highest during the first week of aduhHfe(Greanyetal. 1976). . v w , t, .

PAGE 127

119 Results Immature Development D. longicaudata emergence increased with temperature (Figure 5-1). No parasitoids emerged at 1 1°C, and at 13°C emergence averaged only 22%. No A. suspensa emerged in the control containers (pupae not exposed to parasitoids) at 11°C, and emergence in the controls at 13°C was only 52% of the mean emergence at higher temperatures. For comparison, D. longicaudata emergence at 13°C was 31, 28, 27 and 26% of that at 15, 17, 19 and 2\°C, respectively. Thus, host mortality may be partially, though apparently not totally, responsible for the low parasitoid emergence at low temperatures. Sex ratio (percent males) of the emerging parasitoids increased with temperature (Figure 5-2). Only 3% of the individuals emerging at 13°C were males. Thus, males suffered a higher degree of mortality than females at low temperatures. Temporal distributions of D. longicaudata emergence at various temperatures are illustrated in Figures 5-3 through 5-7. Emergence commenced on days 20, 25, 32 and 49, and peaked on days 24, 30, 39 and 57, for 21, 19, 17 and 15°C, respectively. At 13°C, only 13 individuals emerged between days 78 and 101. The proportion of individuals emerging after the initial peak decreased with temperature (Figure 5-8). At 13°C, 97% of the parasitoids emerged following the transfer of containers from 13 to 25°C (Figure 5-3). At 15°C, mean daily emergence began to increase after ca. 140-150 days, and increased dramatically following transfer to 25°C

PAGE 128

120 (Figure 5-4). Small numbers of parasitoids emerged almost continuously at 17°C (Figure 5-5), while at 21°C most of the late emergence occurred from day 113 (Figure 5-7). Adult Longevity and Ovipositional Activity Adult longevity was inversely related with temperature, averaging 58, 35, 30, 26, 18 and 5 days for females, and 49, 35, 22, 16, 10 and 3 days for males, at 15, 19, 23, 27, 31 and 35°C, respectively (Figure 5-9). Females lived significantly longer than males at 23°C (t=4.20 p<0.0001), 27°C (t=4.88, p<0.0001), 3\°C (t=6.49, p<0.0001), and 35°C (t=7.50, p<0.0001), but not 15°C (t=1.97, p=0.051) or 19°C (t=0.03, p=0.98). A preliminary experiment suggests that mortality does not increase at 13°C, but may increase at 11°C. Four cages with 10 females and 10 males were placed in each of two chambers set at these temperatures for 23 days. Mean mortality for this period was 12.5 and 32.5% for females and males, respectively, at 1 1°C, and 5 and 17% for females and males, respectively, at 13°C, compared to 10% for females and 12% for males at 1 5°C in the preceding experiment. Ovipositional activity as estimated by progeny production on days 4 and 7 was significantly higher in cages placed on the lower shelf at most temperatures (19°C, t=4.38, p<0.0006; 23°C, p=5.70, p<0.0001; 27°C, t=2.68, p<0.02; 31°C, t=4.77, p<0.0003). The relatively low activity in the upper cages may be the result of air movement at the top of the chamber (note that the oviposition unit was placed on top of the cage). Therefore, in comparisons of ovipositional activity among temperatures only data from the cages on the lower shelf were used.

PAGE 129

121 CM tn O) <1) T3 2 0) to Ql ^ E (D CO euednd ;o jsquinu /OujSjaujs spjo;jsejed jsqiuriN ex's CO s >^ "O a. o o o o B 2 S a> O ^ 2 S V § i o C a) 05 C o o M 3 O 2^ *2 CO 9i J3 I S 1 S 00 c o o o at a 53 O (U c CO oi W o ' IS 4) 3 (U C O .«) •4-* o c u I E CO

PAGE 130

122 CNJ o £ O) 0) ' CD 1— 0) a. E c c 2 S S CO tr; *-' rs c c o .2P 3 2 o c CO (L> a> s -£ J5 ssieuj ^uaojed & 5 0 c 1 ° I 2 ^ ^§ ^ .2 5 .2 o 1 ^ CO c*; y! C c S J? s u(U
PAGE 131

123

PAGE 132

124

PAGE 133

125

PAGE 135

127

PAGE 136

128 O) O CO 0) O) 0) -4— » (0 I— a> Q. E H in CO •a £ 'm ^1 >^ o o c £ c (O § «-» tJ ^ s o S CO is S;^ ^ °2 o T3 o 0 V _ i_i C cd > 0) eo.S C en •5b 2 1 § I ^ S u 5P 3 s: CO O ^ o o o I c CO o c 3 Q I 00 G ct] S 3 c o o c/1 3 O 'C > S c a> »-» c a> is i>i it, 5 c >> c op o n u a •*-» -*-» (U s at CO V jC -*-» CO cd CQ

PAGE 137

129 at <^ a> to O

PAGE 138

^Mougf^r-r<-' 130 CM 00 (O
PAGE 139

131 Progeny production was greatest at 27°C (though not significantly higher than at 23°), decreasing with lower and higher temperatures (Figure 5-10). No oviposition was apparent at 35°C. Discussion The late emergence of parasitoids suggests that some individuals undergo arrested development. This may be true diapause (involving induction and release by environmental factors) or non-diapause quiescence (see reviews by Tauber at al. (1983, 1986) for further explanation of these terms). Which of these processes may be occurring in D. longicaudata could not be determined. The phenomenon of arrested development appears to increase at lower temperatures. Ashley et al. (1976), examining effects of high temperatures, observed an increase in delayed development of D. longicaudata larvae at the lowest temperature in their study (22°C). In the present study, delayed emergence was a common occurrence at 15°C, with over 50% of the individuals emerging after the initial peak. At 13°C, almost all parasitoids emerged only following transfer to room temperature. Thus it appears that a polymorphism occurs within the population, where some individuals delay their immature development for various lengths of time. This polymorphism is, however, related to temperature, peaking at 15°C. D. longicaudata emergence decreases with temperature, is very low at 13°C, and no emergence was observed at 1 1°C. However, some of the mortality may be attributed to host mortality. No A. suspensa emerged at 11 °C, which is similar to the 10°C development threshold determined by Prescott and Baranowski (1971). The low fly

PAGE 140

132 i ; • ' emergence at 13°C could account for much, but not all, of the reduced parasitoid emergence at this temperature. , ? Studies with other fruit fly parasitoids have demonstrated somewhat less cold tolerance for the parasitoids than for their hosts. Loni (1997) calculated a slightly higher low temperature threshold for Psyttalia concolor (Szepligeti) than for its host Bactrocera oleae (Gmelin). Psyttalia concolor is apparently somewhat less cold tolerant than D. longicaudata, no parasitoids emerged at 13°C, and emergence at 15°C was less than 1%. However, it is unclear whether nonemerging individuals remained viable at low temperatures. Doryctobracon crawfordi (Viereck) is less tolerant to both low and high temperatures than its host Anastrepha ludens (Loew) (Darby and Kapp 1934). D. crawfordi failed to emerge at 12°C, while emergence for A. ludens at this temperature was 84%. Mortality of D. longicaudata at low temperatures was greater for males than for females. The impact of male mortality on D. longicaudata populations would normally be minimal. However, very low male emergence could resuh in many females not mating and producing only male progeny (=constrained sex allocation, see Godfray 1994). Occurrence of this phenomenon over several consecutive generations could lead to extinction. This is, nonetheless, unlikely under field conditions, as no more than 1-2 consecutive generations would be exposed to sufficiently cold temperatures. The common occurrence of arrested development at 15°C suggests that this species is somewhat adapted to periods of low temperatures. Low adult mortality and reduced ovipositional activity at low temperatures may be additional adaptations to cold conditions.

PAGE 141

133 While there may be some immature parasitoid mortality as a direct effect of low temperatures, it is uncertain whether it is sufficient to explain the absence of D. longicaudata from large areas of the host range in central Florida (Chapter 3), or the low population levels following cold winters (Chapter 4). Loni (1997) concluded that temperature influence alone is insufficient to explain the absence of P. concolor at high latitudes in Italy. The differences in mean minimum winter temperatures between LaBelle, where D. longicaudata is common, and sites to the north where it is absent, are less than 2°C (Table 3-2). The variability among these towns in extreme minimum temperatures is similarly small. Thus small differences in temperature tolerances between the parasitoid and its host may account for absence of the parasitoid. However, it is difficult to relate studies of development in constant temperatures with the dynamic temperature fluctuations typical of field conditions. Extreme temperatures and their frequencies may be important determinants of mortality. Laboratory studies using variable temperature regimens to simulate field conditions are needed to better understand the relationships between temperature and distribution. In order for direct temperature effect to explain the absence of D. longicaudata, this species must be less tolerant to low temperatures than Doryctobracon areolatus (Szepligeti). Therefore, experiments comparing the tolerances of both species are necessary. Unfortunately, difficulties with the maintenance of D. areolatus laboratory cultures have prevented such studies at the present time.

PAGE 142

CHAPTER 6 LABORATORY REARING OF DORYCTOBRACON AREOLATUS Doryctobracon areolatus (Szepligeti) is the dominant native parasitoid of Anastrepha spp. in many areas of the American tropics and subtropics (see references in Chapter 2). However, due to difficulty in rearing this species, it has not been utilized in augmentative release projects. Such releases have been limited to exotic parasitoids easily established in laboratory cultures such as Diachasmimorpha longicaudata (Ashmead) (Sivinski 1996). Other species of opiine braconids successfully reared on tephritid fruit flies in the laboratory include Diachasmimorpha tryoni (Cameron), Fopius arisanus (Sonan), F. vandenboschi (Fullaway) and Psyttalia fletcheri (Silvestri) (Ramadan et al. 1992, 1994b, 1995, Wong and Ramadan 1992). A particularly important problem in rearing D. areolatus has been the dependence on fruits as rearing media for the host fly larvae. Chemical cues emanating from host fruits are apparently essential for host location in this species (Chapter 7). This chapter describes an efficient rearing method utilizing host fruit chemicals to stimulate oviposition into artificial units containing host larvae in diet. Materials and Methods Insects . Parent generation D. areolatus, a total of 128 females and 41 males, were reared out of guava fruits collected at LaBelle, Florida. Host Caribbean fruit fly, Anastrepha 134

PAGE 143

135 suspensa (Loew), larvae were obtained from the mass-rearing facility, Florida Department of Agriculture and Consumer Services, Division of Plant Industry. Cage Setup Adult parasitoids were maintained in 20 cm^ metal-framed cages, the top and two side panels with 16-mesh screens, and other panels Plexiglas. One of the side Plexiglas panels included a cloth sleeve. A brown paper towel was taped to the outside of the opposing Plexiglas panel, in order to cut down on the light intensity. Each cage was stocked over a period of several days (depending of the emergence rate) with approximately 100 females and 100 males. A total of 1, 2 and 5 cages were set up with parent, first and second generation parasitoids, respectively. Parasitoids were supplied daily with a fresh block of honey agar set on an inverted 30 ml plastic cup, and a strip of honey on the Plexiglas side panel, as food. A water dispenser constructed from a 100 ml plastic cup was placed in each cage. Cages were maintained in a room at 25 ± 0.5°C, variable humidity and a 14:10 (L:D) hr photoperiod. Exposure to Host Larvae Oviposition units were composed of A. suspensa larvae in diet (Bums 1995) between two layers of cloth, topped with a layer of parafilm, all maintained within a 7.6 cm diameter plastic embroidery hoop. Prior to exposure, the parafilm had been wrapped overnight on a fresh Anjou pear, previously placed for several hours in a cage with adult A. suspensa. It was placed in the unit exposed side out. Each sheet of parafilm was used

PAGE 144

136 on two consecutive days. Between exposures it was kept in a sealed and refrigerated plastic cup. Approximately 40 cm^ diet containing several hundred host larvae were placed in each oviposition unit. The larvae-diet mixture was selected from areas of the larval trays containing high densities of larvae, so that at least 50% of the volume was larvae. This was done in order increase the chance of successfiil probing by the parasitoids. The amount of diet placed in the unit was considered to be optimal. More diet would have made it too high, allowing larvae to migrate down and avoid parasitism. Less diet would have left parts of the unit devoid of hosts, thus decreasing the chance of a successfiil probing. Host larvae were usually 4 or 5 days old, corresponding to late second and/or early third instar. Occasionally 3 or 6 day old larvae were used. The oviposition unit was placed upon an inverted 100 ml plastic cup. Parasitoids were exposed to hosts for ca. 8 hr daily. However, when high activity (15 or more parasitoids simultaneously on the oviposition unit) was observed, two successive exposures were performed, with units being replaced after 4 hr. This was done in order to reduce the chance of superparasitism. Parasitoids were first exposed to hosts within several days of emergence. Exposure continued daily, depending on availability of suitable hosts. Immature Stages and Adult Emergence Upon completion of exposure, host larvae were transferred to 30 ml plastic cups, which were filled to the top with fresh diet. These cups were then placed upon moist fine vermiculite (15-20 ml water per 100 cm^ vermiculite) in 500 ml plastic cups. Fully developed larvae emerged from the diet, dropping to the vermiculite in which they

PAGE 145

137 pupated. After allowing larvae to emerge for several days, the vermiculite was sieved, and host puparia transferred into fresh moist vermiculite within 100 ml plastic cups. These cups were covered with a solid lid, which was replaced after one week with a cloth lid. This procedure allowed the vermiculite to remain moist while minimizing development of fungi. Immature stages were maintained at the same environmental conditions as adults. Adult parasitoid emergence was recorded daily, and parasitoids transferred to screened cages. Cups were discarded when no emergence was observed for several days. Life History Traits Because cages were stocked over several days, the exact age of ovipositing females could not be determined. Age was estimated as the difference between the exposure date and the median emergence date of all females in the cage. This age estimate was subsequently related with progeny production and sex ratio. Only secondgeneration cages were used in these calculations. At the end of the experiment, the total number of females found dead was only 66 and 65% of the number put in the cages for the first and second generations, respectively. This suggests that a large number of parasitoids escaped. Indeed, some parasitoids were observed pushing through holes in the screen on the top panel of the cage. However, it was assumed that most escaped shortly after being put into the cages, and did not significantly contribute to progeny production. Therefore, the number of females in a cage on a given date was estimated by subtracting the number dying from that day forward from the total dying in the cage.

PAGE 146

138 Results and Discussion Total production and sex ratio of progeny for the three generations are shown in Figure 6-1. Progeny production was very low for the parent generation, with only 2.4 progeny per female. However, the laboratory-reared females produced larger numbers. Production for the first and second generations was 12.1 and 9.3 progeny per female, respectively. This is still much lower than the 29.6 progeny per female reported for D. longicaudata (Greany et al. 1976). Note that the rearing methods differed between these studies, so that this may not be a representative comparison of the reproductive potentials of the two species. The sex ratio of second-generation progeny was highly male-biased (Figure 6-1). Note that these were progeny from females in a single first-generation cage. A similar phenomenon was observed in the progeny of females from a single cage in a previous rearing project. The male-biased sex ratio may be the result of many females not having mated. The cause of this is unclear. Mean daily progeny production was between 1-2 progeny per female for almost all days from age 9 to 22 days (Figure 6-2). In contrast, daily progeny production for D. longicaudata was shown to peak at nearly 4 progeny per female, but remained above one progeny per female for only 9 consecutive days (Greany et al. 1976). Note that for both species the number of mature eggs in the ovaries (D. areolatus, 64.3 ± 4.3, n=6; D. longicaudata, 73.0 ± 6.3, n-6; t=1.18, p=0.26; 7-day-old females not exposed to hosts, specimens obtained from Martin Aluja, Instituto de Ecologia, Xalapa, Veracruz, Mexico) is much greater than the maximum number oviposited per day. The differences in progeny production may be the resuh of different experimental procedures or differential

PAGE 147

139 adaptability to laboratory conditions. However, they may also represent different reproductive strategies, whereby D. longiccmdata produces large numbers of progeny in a short period of time, and D. areolatus smaller numbers over longer periods. If this were the case, it would give young D. longiccmdata females a competitive advantage in exploiting host patches (see discussion in Chapter 3). The sex ratio of progeny of second-generation females was relatively stable over time, averaging close to 50% (Figure 6-3). However, at the oldest female ages the progeny sex ratio tended to be male-biased. This may be the result of sperm depletion, or perhaps lower mortality of unmated females. Immature development time at 25°C was 22.1 ± 1.1 days (range 19-35 days) for females and 20.6 ±1.1 days (range 18-26 days) for males. The rearing method reported here is a vast improvement over the previously used method of rearing on hosts within fruits. While production numbers are lower and costs higher than with D. longicaudata, this method could still serve as a basis for establishment of laboratory cultures for research. Further improvements in rearing techniques could make possible mass-production for purposes such as augmentative releases. Chemical identification of fruit cues used for host location might totally eliminate the need for fruits. Improvements in nutrition and control of pathogens would also be beneficial. Other improvements including smaller mesh screens to prevent escape and differently designed cages with more surface area for resting have already been implemented in subsequent studies (see Chapter 7).

PAGE 148

140

PAGE 149

141

PAGE 150

142

PAGE 151

CHAPTER 7 BEHAVIORAL RESPONSE TO HOST CHEMICAL CUES BY FEMALES OF DORYCTOBRACON AREOLA TVS Diachasmimorpha longicaudata (Ashmead) locates hosts within fmits by sensing the vibrations of the moving larvae (Lawrence 1981). Females of this species respond to an artificial apparatus containing host larvae with immediate landing, probing and oviposition. In contrast, Doryctobracon areolatus (Szepligeti) females show no such response. However, if a sheet of parafilm previously wrapped on a fruit exposed to flies is incorporated into this apparatus, flight to the device and oviposition response is observed. This suggests that host chemical cues are essential in the host location behavior ofZ). areolatus. Chemical cues are important facilitators of host location in many parasitoids. These may be associated with either the host insect or with the plant on which it feeds (Godfray 1994). This study is designed to determine the source of the host cues utilized by D. areolatus, i.e. whether they are associated with the host fly or fruit. Materials and Methods Insects Adult Doryctobracon areolatus were from a third-generation laboratory culture, the parent generation of which was reared from Cattley guava {Psidium cattleiamm Sabine) fruit collected mostly at LaBelle, Florida. Caribbean froiit fly, Artastrepha 143

PAGE 152

144 suspensa (Loew), larvae were obtained from the laboratory colony maintained for ca. 150 generations at the Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Gainesville, Florida. Experimental Design Cages were 30 cm long x 20 cm wide x 20 cm high. The bottom and two longer sides were Plexiglas, with a cloth sleeve in the middle of one of the side panels. The top panel was 52-mesh screen, and the two smaller sides 16-mesh screen. Each of 6 cages was stocked with 100 female and 70 male D. areolatus. Dead females were replaced daily. Oviposition units were composed of several hundred second and/or third instar A. suspensa larvae in diet (Bums 1995) between two layers of cloth, topped with a layer of parafilm, all maintained within a 7.6 cm diameter wooden embroidery hoop. Two oviposition units were placed in each cage upon plastic containers. One unit ("Control") contained parafilm wrapped overnight on unwaxed Anjou pears exposed to ovipositing A. suspensa females for several hours. This was identical to the units used in rearing the laboratory culture (Chapter 6). The second unit contained parafilm with one of the following treatments: (1) "Unpunctured fiuit"wrapped on fresh undamaged pear; (2) "Punctured fruit"wrapped on pear punctured approximately 200x with a no. 0 insect pin (to simulate puncturing by ovipositing flies); (3) "Damaged fiiiit"wrapped on pear from which sections of pulp had been cut out (to simulate vertebrate damage); (4) 'Tly cues"placed for several hours within cage containing ovipositing A. suspensa females (flies oviposited through the parafilm between several to several hundred times); (5) 'Tly cues

PAGE 153

145 + punctured fruit"as treatment (4) but subsequently wrapped on punctured pear; (6) Untreated parafilm. Previous observations indicated that naive individuals (without oviposition experience) showed relatively little response to the oviposition unit. Therefore, prior to experimentation, females were exposed to a control oviposition unit at least once. The experiment was replicated on 12 of 13 consecutive days. On each day, each of the 6 cages contained a different treatment. Each treatment was replicated twice in each cage, once placed on the left side and once on the right side of the cage. The placement on any given day was random. Exposure to each pair of oviposition units lasted 8-10 hr. Fruit fly larvae were transferred to a small plastic container, which was filled with fresh diet. This container was placed within a larger container with moist fine vermiculite (15-20 ml water per 100 cm vermiculite). Mature larvae exited the diet and pupated in the vermiculite. Containers with vermiculite were covered with a solid lid for one week, and subsequently with a screened lid until parasitoid emergence. Response Variables and Statistical Analysis The number of females active on each oviposition unit was recorded at 1 hr, 4 hr, and 8 hr following placement of the units in the cage. An active female was defined as an individual either probing into the unit with its ovipositor, or one standing on the unit with ovipositor at a horizontal or below horizontal position; otherwise the ovipositor is curved slightly upward. The difference between the "Control" and "Treatment" units was calculated for each cage at each hour. The number of D. areolatus progeny from each

PAGE 154

146 oviposition unit was counted, and the difference between the "Control" and "Treatment" units calculated for each cage. Treatments were compared by paired t-test. Results The unpunctured and punctured fruit treatments did not differ from the control (fruit exposed to flies) in terms of number of active D. areolatus females on the oviposition unit (t=1.77, p=0.08; t=0.26, p=0.80; respectively). The fly cues + punctured fruit, damaged fruit, fly cues, and untreated parafilm treatments all showed less response than the control (t=2.24, p=0.03; t=3.83, p=0.0003; t=7.28, p=0.0001; t=6.80, p=0.0001; respectively). Figure 7-1 compares the response of females among the various treatments. All of the fruit treatments showed significantly greater response that either fly cues or untreated parafilm. Additionally, response to punctured fruit odor was greater than that to odor of damaged fruit. In the fly cues + unpunctured fruit treatment, there was a significant reduction in the number of females on the oviposition unit after 8 hr, as compared to both 1 hr and 4 hr (t=4.23. p=0.0001; t=2.43, p=0.01; respectively). Consequently, there was a significantly lesser response after 8 hr to this treatment than to either the unpunctured or punctured fiiiit treatments (t=2.50, p=0.04; t=3.67, p=0.002; respectively). The cause of this is unclear. The number of progeny emerging from the unpunctured fioiit, punctured fruit and fly cues + punctured fioiit treatments was not significantly different than the number emerging from the control units (t=0.07, p=0.94; t=1.48, p=0.14; t=2.24, p=0.03; respectively). Progeny emergence from the damaged fruit, fly cues and untreated parafilm treatments was significantly lower than from the control units (t=2.66, p=0.01;

PAGE 155

147 t=4.67, p=0.0001; t=5.00, p=0.0001, respectively). Figure 7-2 compares the emergence among the various treatments. Emergence from the fly cues and untreated parafilm treatments was significantly lower than from all fruit treatments except damaged fruit. There were no significant diflFerences in emergence among fruit treatments. Discussion Several studies have demonstrated the importance of host-associated chemicals in host location by parasitoids of fruit flies. Greany et al. (1977) found that chemicals released by fiingi associated with rotten fruits are attractive to D. longicaudata females. Messing and Jang (1992), using chopped ripe fruits placed in traps, demonstrated attraction of D. longicaudata females to various host fruits. Messing et al. (1996) demonstrated similar responses by Psyttalia fletcheri (Silvestri) to odors of fresh cucumber and decaying pumpkin. These studies measured adult attraction and only infer that this response is related to oviposition. The current study differs in that the response variables directly measure ovipositional activity. This study clearly demonstrates the importance of chemical cues emanating from ripe host fruit in the host location behavior of D. areolatus. Response to unpunctured fruit did not differ from that to punctured fruit. This suggests that response is unrelated to host fly ovipositional activity, which the puncturing was designed to simulate. Response to damaged fruit was somewhat less than to the other fruit treatments. As the peel was removed and the pulp exposed in this treatment, this suggests that the active chemical(s) may be located in the peel of the fruit. The bioassay used in this study could not determine if the fruit chemical(s) act as attractants, arrestants, or oviposition stimulants, or if they are volatile or contact

PAGE 156

148 CO
O o II £3 wJ u. £ «5 (U O ti O ° •2 « E ^ 3 4) O c c lO CO CM JUaUJlBQJJ UO S8|BUJ8i lOJJUOO UO S9|BLU9d OS ^ o 3 CQ S C CO 5 (u c £ c ^ C ea a> (u ^ <» o c o fll •-• 2 ^ O ^ o

PAGE 157

li * 149 3 W IL»

PAGE 158

150 chemicals. More than one class of chemicals may be involved, or one chemical may have a role in successive stages of the host-location behavior. Females have been observed landing on the periphery of the oviposition unit, and walking onto the parafilm while antennating. This behavior suggests that volatile fruit chemicals may be at least partially involved in host location. Prokopy and Webster (1978) found that Utetes canaliculatus (=Opius lectus) (Gahan) responds primarily to the host-marking pheromone of Rhagoletis pomonella (Walsh). Similarly, Halticoptera rosae Burks (Hymenoptera: Pteromalidae) was found to respond to the pheromone deposited by Rhalogelis basiola (Osten Sacken) (Roitberg and Lalonde 1991). Chemical cues derived from the host fly have no apparent effect on D. areolatus females. While both U. canaliculatus and H. rosae parasitize eggs or earlyinstar larvae, D. areolatus prefers later instars (Chapter 6). As the host pheromone is water-soluble, it would be degraded by precipitation. Thus the former two species, which attack the host shortly following oviposition, should conceivably have a closer association with it, and more likely evolve a response. D. longicaudata can locate larvae within fruit solely by vibration sensing (Lawrence 1981). Vibrotaxis has also been reported for Diachasmimorpha mellea (Gahan) (Lathrop and Newton 1933), Diachasma alloeum (Muesebeck) (Glas and Vet 1983), and Aganaspis pelleranoi (Brethes) (Hymenoptera: Eucoilidae) (Ovruski 1994). Henneman (1996) demonstrated that Diachasmimorpha juglandis (Muesebeck) females discriminate between infested and uninfested walnut fruits before landing, but did not determine which cues may be involved. D. areolatus requires chemical cues associated with the fruit for host location. However, this does not imply that vibration cues are not

PAGE 159

151 used by this species. Chemical cues may be used in the early stages of host location, as attractants or arrestants, with vibration stimulating probing behavior once the parasitoid is on the fruit. The greater dependence of D. areolatus on host fruit odors relative to D. longicaudata suggests a greater affmity to these odors, and perhaps an advantage in locating host patches. Distribution patterns of these two species in the field are consistent with D. areolatus being superior to D. longicaudata in this regard (Chapters 3 and 4, Sivinski et al. 1998).

PAGE 160

CHAPTER 8 SUMMARY AND CONCLUSIONS Two main trends were observed in the distribution of Caribbean fruit fly {Anastrepha suspensa (Loew)) parasitoids in Florida (Chapter 3). First, D. longicaudata was absent from a large area of interior central Florida, north of the towns of LaBelle and Okeechobee. Second, D. areolatus was absent from the Atlantic coast and rare on the coast of the Gulf of Mexico. Populations of both D. longicaudata and D. areolatus are reduced following cold winters. Nevertheless, D. longicaudata is more adversely affected (Chapter 4). Similarly, the geographic region in which D. longicaudata is absent is characterized by low winter temperatures. However, absence of D. longicaudata is best explained by high variability in temperatures. Two hypotheses are proposed to explain the effect of temperature on D. longicaudata. (1) Low temperatures have a direct negative effect. (2) Low or variable temperatures resuh in periods of low host availability, and such temporal gaps in hosts are detrimental to D. longicaudata survival. Field data from Florida are generally consistent with both hypotheses. Laboratory studies using constant temperature regimens suggest that D. longicaudata is somewhat less tolerant to low temperatures than is the host A. suspensa (Chapter 5). However, it is unclear whether this difference is sufficient to explain absence of the parasitoid from large regions of the host's range. Additionally, the phenomenon of 152

PAGE 161

153 delayed emergence of D. longicaudata at low temperatures suggests a certain adaptation to these conditions. Several lines of evidence support the hypothesis of an indirect temperature effect on D. longicaudata distribution. First, variances of monthly temperatures, rather than absolute temperatures, are the most significant variables related with presence at a certain site. Variability of temperature among months is a measure of seasonality, and is likely to be related with host availability. Furthermore, a positive relationship was observed between D. longicaudata abundance and numbers of host flies captured in traps at various sites, while no such relationship was found for D. areolatus. This suggests that D. longicaudata is less successful under conditions of low host density. Such conditions would be more frequent at locations with greater seasonal fluctuations in temperature. The rare occurrence of D. areolatus along the coasts, and especially its absence in the area of its original introduction, may be the resuh of interspecific competition. There is little overlap in distribution with D. longicaudata, which was introduced three years later. Of all the sites surveyed, only at LaBelle do both species co-occur in large numbers. Studies at LaBelle suggest that significant competition may occur at least at the end of the Surinam cherry fruiting season (Chapter 4). The negative among-site relationship in abundance between D. areolatus and U. anastrephae (Chapter 3) may also be indicative of competition. Distribution patterns at LaBelle are consistent with "counter-balanced competition" (cf Zwolfer 1971) where D. areolatus is superior to D. longicaudata in locating host patches (=extrinsic competitor) and D. longicaudata is superior in exploiting these patches (=intrinsic competitor) (Chapter 4, Sivinski et al. 1998). Host

PAGE 162

154 fruit chemicals are important cues in the host location behavior of D. areolatus, and are apparently essential for stimulating oviposition in the laboratory (Chapters 6 and 7). This may be related to its ability to locate host patches. Conversely, D. longicaudata does not require fruit cues to locate host larvae in the laboratory, and it has been shown to locate hosts within fruits using vibration cues (Lawrence 1981). This may be related to its ability to exploit host patches. Note that attraction to fresh host fruit chemicals has been shown for D. longicaudata in the field, and it is unclear whether D. areolatus can detect vibrations of host larvae. Further studies comparing these two species are needed to resolve how they may be differ in host and host-habitat location strategies. Several additional mechanisms may contributed to a competitive advantage for D. longicaudata over D. areolatus. Perhaps the most obvious is its longer ovipositor, which enables it to reach host larvae deeper within the fi\iit. Comparisons of laboratory studies suggest that D. longicaudata females may lay larger numbers of eggs during the first days of adult life (Chapter 6, Greany et al. 1976). If this is representative of behavior in the field (which is far from certain), it would enable faster exploitation of host patches. Finally, studies have indicated that at least under certain circumstances, D. longicaudata larvae may have an advantage in physical competition with other parasitoid species within host larvae (see references in Chapter 3). Further studies are needed to determine whether such an advantage exists over D. areolatus larvae.

PAGE 163

APPENDIX NUMBERS OF SAMPLES COLLECTED AND INSECTS EMERGING FOR VARIOUS SITES BY MONTH, YEAR AND FRUIT TYPE

PAGE 164

156 •o Q PL, u +-> -o ^5 X! ^ Q Ml CO o H ooooo<^oooooooo OOOOOOOOOOOOCNO 00 00 Tf m m oo o oo r--o o m 00 00 00 00 m fS (S O CO §) e o S s o U ^ ^ b l-< u o X! o CO 3 o CO 3 cr o iri m ON ON On On On On CO 3 (T O OS On CO 3 o in On On 6 CO c •c 3 3 C« i/3 e CO c X! o CO > CO CO > CO 1 00 (J O On ON in ON ON m On On V) 3 6X) 2 •3 CO o CO 3 C CO CO 2 x> a> x: CO CO 2 < S 1^ .3 .3 in ON ON CO 3 o o o o o o o' o CO CO CO > > > CO en CO 3 00 3 00 e c e o o o e S S s £ s o o o U U U
PAGE 165

157 CO c <1> 6 u CO o e I a o (1> JO S 3 J? E 2 J3 C o Ph ^5 000000'<4-00(N0-^0-^0(^00 0000'-^0(N^O^O(SOO^O^O oooooooooooooooooo OOOOOO^OO— §) 1 ON PQ > §) c o s E o o in ON 2 > g. c o s £ _ o (J c o E E o GO O o E c c 3 3 cr o h-) o s c •c 3 o E E o (L> ej E ed c c 3 c<3 > c o E E o s c •c 3 c4 > §) c o E E o u E «J c •c 3 > eS 3 00 c o E E o in §5 "3 >n ON On (U E G c 3 m ON On v x> E a. a> C/3

PAGE 166

CO c e S 3 Cm O (U S T3 E s c o J3 Q U ^5 •r CO ^ Q O H OOOOOO-— 'OOOOOOOr^OOO vOOOOOOOOOOOOOOVOOOO 000000000000000000 O VO > 3 o E E o u E -4-» a. 00 o >0 iri — m o On ir> vo ^ «r> 000000 — OOOOOOOmOOO CSOOOOOOOOOOOOO-^OOO 000000000000000000 > 3 bO 00 o E E o u > CO CO ON 0\ ON ON cd 3 3^ cr cr C O O cS J J u On on On on 5 CO 3 ao c o E E o u CO 3 o E CO c •c 3 0) x: o E CO c •c 3 CO > CO c o o E CO c C/5 (/3 E O 3 CO 3 o CO 3 o o E CO c •c 3 <1> o E CO c c 3 CO > CO c o E E o NO NO ON On On On On On E NO >/-l i/N On ON On On On ON On On (U C/3 CO < >^ CO 3) o 5 i I CO rJ*

PAGE 167

159 00 e 'So s o o u U S 1/1 e CO o B 3 at CO H PL, o CO Q •r OS ^ Q +-» o H u-i 'aOOOOCSOOOfn-^CSt-H.-HOOOOO On OOOOCSOOOO^'— ir<^(SOCSro>r>0 OOOOOOOOOOOOOOO-— 0 3 > cd 3 3 cr o H-1 CO 3 cr o c o s E o c o E E o 00 > 03 3 eso e o E E o 00 ^ ^ VO m o E c •c 3 > 3 60 c o E E o vo m (N ^ r-l o ^ o fN — XI X o o o E E E E to CO to CO CO CO 3 rin 3 c rin c cr •c •c o 3 o 3 3 3 hJ 05 1/3 Vi ir> r> IT) TjOn On On OS <7\ On On On On ON X CJ u CO a. < (A 3 60 X E (U a. ON On CO 3 C CO On On CO 2 X u ON On X o ir> ID On on On On in ON ON CO > CO 3 60 C o E E o CO > CO CO ON On ON On CO > CO 3 60 C o E E o U T3 CO Q r3 ^ 3 3 t; >> 60 60 CO 3 CT O h-) ON ON 3 C CO

PAGE 168

160 'ijri
ON (S Oirir-Hrn^^O'rr'^'^'^ONO'OOOOO \0 CN oot^c^or^-oo^'^ONm'^'— 'CNOooo O^OOCSOO-^tT^OOOOOOOO ^(^ON'-^OON'Oiri'— w-ivocNr^c^oOONvOOO O'«t-CS(NTl-^OCSf«^'- CO CO > CO 3) CO > CO CO > CO "5 CO c o E E o (J U U CO CO CO 3 (T O u o E CO c •c 3 CO > CO 3 £S0 o E E o U CO C iTi iT) ON On ON On ON On 2 J3 -3 ^ i a> o E CO c •c 3 ON ON CO CO 3 cr o CO 3 O" O CO 3 O" O CO 3 CT O >0 l/N »n IT) On ON ON On 0\ On On On CO 3 C CO u CO 2 X) o CO

PAGE 169

161 60 c s u CO U o \-, E I CO B 3 T3 S CO H CO c o CO CO Q J3 •kT CC •r CO CO oooooooooot^ooooooo (N On 0\ ON OOOO«rv00'-<^00 (L> o s CO c 0\ — ^ o o r~00 ^ O ^ CN --H ON Tf Tf rj^ ^ O — ' — ' ONOOsTi-cnoo t~~ o 00 ooocNOi/itscNu-i^rtr-cNtS'-^f^^mo o o o o o O ^ O O —I m m /-i »n OS On On On ON On a, ,^ S < S ^ U C CO > CO 3 at) c o E E o U ON ON On On 4> CN m >0 o o o CS ^ —1 (S o '—' m n >o w-i ON ON ON On ON ON CO > CO 1 CO > CO 3 00 c o E E o CO > CO 3 00 1 CO > CO 3 00 e o E E o
PAGE 170

162 00 c s •*-* O a> Vi o E J? S o l-c X) 3 T3 E at (/> 2 i3 c o •a CO 0} Q PL, u •r ee ^5 ^ Q cd *-» O H OOOOOOOOOOOfSOOOOOO 0000000000000000000 o 00000000>/^ON\OmfN'-<-^ VO On o — r00 ocscs^<^r^TroNTi-r~'* — ovoo Tj-T-^'O'^'— ivOvO'— iriO\t--vOOS'— OOOOOOOOOOO'-OOOOOO CNOOOOOOOOOOOOOOOOO CSr^OOOO^fNCS'— 'tSfSt-Hr-omcSvO o 00 Ov^r'^r~t-~'^<^rrvof<^'ooooTfvo^(SON^'- £0 > OS 3) ^ c o E E o c o E E o CO 1 c« (J in m On On ON ON X) E tn 00 < (U c — CO I3 C CO CO 5 PL, Si o CO CO S 1 CO CO 00 > > > CO CO CO a 3 00 c c c 0 0 0 E E E E E E 0 0 0 CJ u u m ON On ON On 1 E CA 3 00 3 (/3 < to 3 3 3 cr (T cr 0 0 0 J in m m ON ON ON ON ON ON CO 3 cr o h4 x; S CO c •c 3 on u-N m ON ON ON ON CO 3 e cO 2 x: o CO

PAGE 171

163 C £3) i-i s CO --» O a> c E a, S CO E "a. E 2 4-» c o 5 CO Q J3 J3 ^ Q •5 O H '-^OOOOOOOOOOOOOOOOO oooooooooooooooooo o E ca c •c 3 in 0^ —< o 00 00 C3 > at 3 ON 0\ ,3 rU-1 o o cs OS o o O o o o ^1 O OS > > > CO a §) 3) 3 00 c c n o o o E E E E E S o o o u O U >n OS OS OS On OS OS ii a> +-> CO E -^ t/5 3 00 *-» cx ^ \o — — ' o o o o o —I so (N (S ^ o r(N 00 (N r<-i O 0\ ^ o OS OS ^ ^ ^ ^ ^ (U 3 o 3 CT O in OS OS 3 o in OS OS E C •c 3 E c c 3 m OS OS o E OS c •c 3 3 00 c o E E o at > at 3 00 c o E E o 3 cr o 3 O H-1 at 3 a" o 4> E CO c •c 3 O E CO c •c 3 m m m OS OS OS OS OS OS m m m Os Os Os OS OS 0\ C/3 C/5 in OS OS at 2 J3 CO c 3 CO 3 00 4> E (L> +-» 3 C at CO 2 J3 O CO CO I

PAGE 172

164 00 c e o 0) c 0) Xi s I as o S e CO <» *-» 2 CO a> c o Q -i-> -a Q O H OOOOOOOOOOOOOtNOOOO OOOOOOOOOOOO^fNOOOO oooooooooooooooooo OOOOOOOOOOOOO"— 3 ao >> > CO 3 60 C O S S o U X! O s CO c c 3 C/3 CO > CO >> 1 CO > CO 3 00 c o S S o CO 3 cr o CO 3 o I o E CO c 3 ir> in 0\ On :5 ^ On on CO 3 cr o h-1 ON On CO o s CO _c 'C 3 CO CO > CO 3 (50 c o S E o ON ON 3 00 CO 3 CO 3 o o u O E CO c •c 3 CO > CO 3 00 c o E E o V5 U On ON CO 3 C CO ON ON CO 2 CO 3 o ON I £ 3 o

PAGE 173

165 as Q J3 ^5 XI 4-1 i ^ Q o H OOt--r-HOOOOOOO 0\ or-fioo^r^'sroooo S 00 O O — I "O O CO §) >> 1 > 3 CX) «/-> >r> > > CO 3 SiO 3) a c c c o o o B B £ B B B 3 o o o o O in ON ON On ON as On o r-H TjOn ^ CO > CO NO O O On r-i ON ^ a> CO 3 o h4 CO 3 CT O ON ON On on CO 3 cr o On On m O s CO CO 3) 3) (U (U i C/5 (/3 C/3 U CJ IT) in tn >n On On On On On On On On < S ^ ^ 3 X5 6 CO b CO 3 C CO CO I = X! o < CO 2 .3 3 3) 3 <

PAGE 174

166 Q X! •r CO •r CO ^ Q O H ooooooooooooooooot^ OOO^^OOO'OOOOOOOOO—' On ^ OOOmOOOOfNlOOOOOOOO— 1 O On O O O O > CO — u ON ON E C/3 "a. NO CS NO ^ O ^ ON 00 OOOOOOOOOOOOOOOOO— ' OOO^^OOO'-'OOOOOOOO O On «D TjCO > CO §) c o S S o U On ON CO 3 SO c o E E o C/3 U CO 3 cr o IT) 85 u E CO c c 3 CO CO 3 o §5 x; o E CO c •c 3 CO > CO 3 CO 3 SO c o E E o (/) CO U 85 CO 3 o >> fc (U x: x: u o E E to CO CO 3 c c O" •c •c O 3 3 cn CO NO NO ON On On On CO X5 u CO

PAGE 175

167 ^5 ^5 ^ Q -»-> Ml H OOOOf^OOOOOOOOOOOOOO oooooooo^ooooooooo OOOOOOOOCSO^OOOOOOOOO •^CNONr0000 O OCN I' ^ ^ ^ ^ V a a a i> 03 3 o H-1 d a* o o S C •c 3 C/3 On 0\ OS ON 3 CT O OS 0\ (U o s c •c 3 s CO G •c 3 O s c •c 3 O E c •c 3 > 1 C/3 C/) (/3 (/5 U «n in ON ON ON ON 2 •a CO 3 Oh l-l 03 C 3 OS ON "3 (1> o e C •c 3 ON ON 05 a > at 3 SO > (0 3 (50 c o S 6 o (J U On On o o o O O O o o o o o O o o o o o O o o
PAGE 176

168 :3 CO Q Pi ^5 •r CO ^ Q o ^OOOOOOOOOOOOOOOOO oooooooooooooooooo oooooooooooo-^oo — oo (S On ON O O >— I ^v^-^voooomcNvoocN>rir-csw->r^\0(S CNOOOOOOOOOOOOOOOOO oooooooooooooooooo oooooooooooooooooo o o o o o o H c •c 3 3 3 C/3 CO > CO at > CO eO > CO 3 00 c o E s o 00 U c o E E o o u u c o E CO 3 o CO 3 o CO 3 cr o I c o E E o u in in On On ON ON m ON ON E c/3 OS ON 3 00 >n m >n On On On On On On CO 3 cr o m On On O E CO c •c 3 at > CO S) c o E E o O E CO e or) U c/D CO 3 00 c o E E o U \0 rt On on On On »-» CO +-* CO *-» CO to 4-t CO 3 3 3 3 3 O" cr CT cr cr O o O o o hJ m m m m m On OS On ON On On ON ON ON On CO 3 C CO CO 2 J3 3 00 CO CO 3 C CO CO 2 o CO I GO CO 2 X) o I

PAGE 177

169 c S (i> 1/1 --» o 0> CO _C O k> I £ O E 3 •o S CO 2 J3 o T3 :3 J3 •kT CO ^ Q o H vO^O^m'^OOOOOO^^OOOO oooooooooooooooooo O CN in in o --I On CS m ^ fi-i m (s m erry erry erry erry x: o o s E «-» E E 0] CO cd CO CO C C 3 3 c c •c •c O" O" •c •c 3 3 O O 3 3 WO hJ >n >n NO NO NO ON On ON ON On ON On ON On ON o NO ^ CO > CO 3 ^ o o — ' ^ r-i On — I o o — ' On VO CO c o E E o CO > CO §) u 1 NO CO > CO 3 ao c o E E o »n m in m ON ON On On On On On On C/) C/) CJ (J (J in «n m m in On On On On On On On On On On CO o. < CO CO 3 GO 1 I 3 2 J3 O eO < CO 2 c 3 .3 3 3 <: E a>

PAGE 178

170 00 c (U s m 0000 ^in in o\ o\ o\ ON o 6 c as > 3 ^ ^ ^ c o S 3 O C/5 U 3 cr o a> O s C o s c c -c 3 3 x: o s c •c 3 CO > CO CO > 03 eO > CO 3 ao c o B S o C/5 (>0 V3 U O (J w~i >r> in in 0\ ON ON ON On ON On On 3 C CO x: o CO CQ CO CO 2 S ^ 3 00

PAGE 179

REFERENCES CITED Aguiar-Menezes, E. L. and E. B. Menezes. 1997. Natural occurrence of parasitoids of Anastrepha spp. Schiner, 1868 (Diptera: Tephritidae) in different host plants, in Itaguai (RJ), Brazil. Biol. Control 8: 1-6. Aguiar, E. L. and E. B. Menezes. 1996. Population dynamics of fruit flies in Itaguai County, State of Rio de Janeiro, Brazil. ISurvey of species, pp. 321-323. In B. A. McPheron and G. J. Steck (eds.). Fruit flies: A world assessment of their biology and management, St. Lucie Press, Delray Beach, Florida. Aluja, M. 1994. Bionomics and management o^ Anastrepha. Annu. Rev. Entomol. 39: 155-178. Aluja, M., M. Cabrera, E. Rios, J. Guillen, H. Celedonio, J. Hendrichs and P. Liedo. 1987. A survey of the economically important fruit flies (Diptera: Tephritidae) present in Chiapas and a few other fruit growing regions in Mexico. Fla. Entomol. 70: 320-329. Aluja, M., J. Guillen, P. Liedo, M. Cabreta, E. Rios, G. de la Rosa, H. Celedanio and D. Mota. 1990. Fruit infesting tephritids (Dipt.: Tephritidae) and associated parasitoids in Chiapas, Mexico. Entomophaga 35: 39-48. Aluja, M., H. Celedonio-Hurtado, P. Liedo, M. Cabrera, F. Castillo, J. Guillen and E. Rios. 1996. Seasonal population fluctuations and ecological implications for management of Anastrepha fruit flies (Diptera: Tephritidae) in commercial mango orchards in southern Mexico. J. Econ. Entomol. 89: 654-667. Aluja, M., M. Lopez and J. Sivinski. Submitted. Evidence for diapause in four native species of larval-pupal fruit fly (Diptera: Tephritidae) parasitoids in tropical environments. Ann. Entomol. Soc. Am. Araujo, E. L., R. A. Zucchi and N. A. Canal D. 1996. Caracterizacao e ocorrencia de Anastrepha zenildae Zucchi (Diptera: Tephritidae) e seus parasitoides (Hymenoptera: Braconidae) numa nova planta hospedeira, no Rio grande do Norte. An. Soc. Entomol Brazil 25: 147-150. Ashley, T. R., P. D. Greany and D. Chambers. 1976. Adult emergence in Biosteres {Opius) longicaudatus and Anastrepha suspensa in relation to temperature and moisture concentration of the medium. Fla. Entomol. 59: 391-396. 171

PAGE 180

Back, E. A. and C. E. Pemberton. 1918. The Mediterranean fruit fly in Hawaii. U.S. Dep. Agric. Bull. 536, 118 pp. Baranowski, R. M. and R. W. Swanson. 1970. Introduction of Parachasma (= Opius) cereus (Hymenoptera: Braconidae) into Florida as a parasite of Anastrepha suspensa (Diptera: Tephritidae). Fla. Entomol. 53: 161-162. Baranowski, R., H. Glenn and J. Sivinski. 1993. Biological control of the Caribbean fruit fly (Diptera: Tephritidae). Fla. Entomol. 76: 245-251. Bateman, M. A. 1972. The ecology of fruit flies. Annu. Rev. Entomol. 17: 493-518. Bennett, F. D., M. Yaseen, M. N. Beg and M. J. Sommeijer. 1977. Anastrepha spp Investigations of their natural enemies and establishment of Biosteres longicaudatus in Trinidad, West Indies. Commonweahh Institute of Biological Control Tech. Bull. 18: 112. Bautista, R. C. and E. J. Harris. 1997. Effects of multiparasitism on the behavior and progeny development of oriental fruit fly parasitoids (Hymenoptera: Braconidae). J. Econ. Entomol. 90: 757-764. Bess, H. A. 1953. Status of Ceratitis capitata in Hawaii following the introduction of Dacus dorsalis and its parasites. Proc. Hawaii. Entomol. Soc. 15: 221-234. Bess, H. A., R. van den Bosch and F. H. Haramoto. 1950. Progress and status of two recently introduced parasites of the oriental fruit fly, Dacus dorsalis Hendel, in Hawaii. Proc. Hawaii. Entomol. Soc. 14: 29-33. Bess, H. A., R. van den Bosch and F. H. Haramoto. 1961. Fruit fly parasites and their activities in Hawaii. Proc. Hawaii. Entomol. Soc. 17: 367-378. Brenner, R. J., D. A. Focks, R. T. Arbogast, D. K. Weaver and D. Shuman. 1998. Practical use of spatial analysis in precision targeting for integrated pest management. Am. Entomol. In Press. Bums, R. E. (ed ). 1995. Procedures manual for mass rearing the Caribbean fruit fly Anastrepha suspensa (Loew) (Diptera: Tephritidae). Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Gainesville, Florida. Bums, R. E., J. D. Diaz and T. C. Holler. 1996. Inundative release of the parasitoid Diachasmimorpha longicaudata for the control of the Caribbean fruit fly, Anastrepha suspensa, pp. 377-381. In B. A. McPheron and G. J. Steck (eds ), Fmit flies: A worid assessment of their biology and management. St. Lucie Press, Delray Beach, Florida. Canal D., N. A, R. A. Zucchi, N. M. da Silva and F. L. Leonel Jr. 1994. Reconocimiento de las especies de parasitoides (Hym.: Braconidae) de moscas de las frutas (Dip.:

PAGE 181

173 Tephritidae) en dos municipios del Estado de Amazonas, Brasil. Boletin del Museo de Entomologia de la Universidad del Valle 2: 1-17. Canal D., N. A., R. A. Zucchi, N. M. da Silva and S. Silvera Neto. 1995. Analise fauistica dos parasitoides (Hymenoptera, Braconidae) de Anastrepha spp. (Diptera, Tephritidae) em Manaus e Iranduba, Estado do Amazonas. Acta Amazonica 25; 235-246. Celedonio-Hurtado, H., M. Aluja and P. Liedo. 1995. Adult population fluctuations of Anastrepha species (Diptera: Tephritidae) in tropical orchard habitats of Chiapas, Mexico. Environ. Entomol. 24: 861-869. Christenson, L. D. & R. H. Foote. 1960. Biology of fruit flies. Annu. Rev. Entomol. 5: 171-192. Clausen, C. P. 1978. Tephritidae, pp. 320-335. In CP. Clausen (ed ). Introduced parasites and predators of arthropod pests and weeds: A world review. USDA Agric. Handbook 480. Clausen, C. P., D. W. Clancy and Q. C. Chock. 1965. Biological control of the oriental fruit fly (Dacus dorsalis Hendel) and other fruit flies in Hawaii. USDA Tech. Bull. 1322, 102 pp. DeBach, P. and D. Rosen. 1991. Biological control by natural enemies. Cambridge University Press, Cambridge. 440 pp. Darby, H. H. and E. M. Kapp. 1934. Studies on the Mexican fruit fly, Anastrepha ludens (Loew). USDA Tech. Bull. 444, 20 pp. Eskafi, F. M. 1990. Parasitism of fruit flies Ceratitis capitata and Anastrepha spp. (Diptera: Tephritidae) in Guatemala. Entomophaga 35: 355-362. Fehn, L. M. 1982. Influencia dos fatores meteorologicos na flutuagao e dinamica de popula<;ao Anastrepha spp. Pesq. Agropec. Bras. 17: 533-544. Fernald, E. A. 1981. Atlas of Florida. Florida State University Foundation, Tallahassee, Florida. 276 pp. Fernandez de Araoz, D. and A. J. Nasca. 1984. Especies de Braconidae (Hymenoptera: Ichneumonoidea) parasitoides de moscas de los frutos (Diptera: Tephritidae) colectados en la provincia de Tucuman (Argentina). CIRPON (Argentina) 2: 37-46. Flitters, N. E. and P. S. Messenger. 1965. USDA Tech. Bull. 1330, 36 pp.

PAGE 182

174 Gilstrap, F. E. and W. G. Hart. 1987. Biological control of the Mediterranean fruit fly in the United States and Central America. U.S. Dept. Agric. ARS-56, 64 pp. Glas, P. C. G. and L. E. M. Vet. 1983. Host-habitat location and host location by Diachasma alloeum Muesebeck (Hym.; Braconidae), a parasitoid of Rhagoletis pomonella Walsh (Dipt., Tephritidae). Netherlands J. Zool. 33: 41-54. Godfray, H. C. J. 1994. Parasitoids. Behavioral and evolutionary ecology. Princeton University Press. Princeton, New Jersey. 473 pp. Gonzalez-Hernandez, A. and L. O. Tejada. 1979. Fluctuacion de la poblacion de Anastrepha ludens (Loew) y de sus enemigos naturales en Sargentia gregii S. Watts. Folia Entomol. Mex. 41: 49-60. Greany, P. D., T. R. Ashley, R. M. Baranowski and D. L. Chambers. 1976. Rearing and life history studies on Biosteres (Opius) longicaudatus (Hymenoptera: Braconidae). Entomophaga 2 1 : 207-2 1 5. Greany, P.D., J.H. Tumlinson, D.L. Chambers and G.M. Boush. 1977. Chemically mediated host finding by Biosteres (Opius) logicaudatus, a parasitoid of tephritid fruit fly larvae. J. Chem. Ecol. 3: 189-195. Haramoto, F. H. and H. A. Bess. 1970. Recent studies on the abundance of the oriental and Mediterranean fruit flies and the status of their parasites. Proc. Hawaii. Entomol. Soc. 20: 551-556. Harris, E. J., J. M. Takara and T. Nishida. 1986. Distribution of the melon fly, Dacus cucurbitae (Diptera: Tephritidae), and host plants on Kauai, Hawaiian islands. Environ. Entomol. 15: 488-493. Harris, E. J. and C. Y. L. Lee. 1986. Seasonal and annual occurrence of Mediterranean fruit flies (Diptera: Tephritidae) in Makaha and Waianae valleys, Oahu, Hawaii. Environ. Entomol. 15: 507-512. Harris, E. J. and C. Y. L. Lee. 1987. Seasonal and annual distribution of the Mediterranean fruit fly (Diptera: Tephritidae) in Honolulu and suburban areas of Oahu, Hawaii. Environ. Entomol. 16: 1273-1282. Harris, E. J. and C. Y. L. Lee. 1989. Influence of bittermelon, Momordica charantia L. (Cucurbitaceae), on distribution of melon fly, Dacus cucurbitae Coquillet (Diptera: Tephritidae) on the island of Molokai, Hawaii. Proc Hawaii. Entomol. Soc. 29: 49-56. Henneman, M. L. 1996. Host location by the parasitic wasp Biosteres juglandis (Hymenoptera: Braconidae) under field and greenhouse conditions. J. Kansas Entomol Soc. 69: 76-84.

PAGE 183

175 Hennessey, M. K. 1994. Analysis of Caribbean fruit fly (Diptera: Tephritidae) trapping data, Dade County, Florida, 1987-1991. Fla. Entomol. 77: 126-135. Hernandez-Ortiz, V., R. PerezAlonso and R. A. Wharton. 1994. Native parasitoids associated with the genus Anastrepha (Dipt.: Tephritidae) in Los Tuxtlas, Veracruz, Mexico. Entomophaga 39: 171-178. Houston, W. W. K. 1981 . Fluctuations in numbers and the significance of the sex ratio of the Mexican fruit fly, Anastrepha ludens caught in McPhail traps. Entomol. Exp. Appl. 30: 140-150. JimenezJimenez, E. 1956. Las moscas de la fruta y sus enemigos naturales. Fitofilo 16: 4-11. JimenezJimenez, E. 1958. Empleo de enemigos naturales para el control de insectos que consituyen plagas agricolas en la Repiiblica Mexicana. Fitofilo 21: 5-30. Jiron, L. F. and I. Hedstrom. 1988. Occurrence of fruit flies of the genera Anastrepha and Ceratitis (Diptera: Tephritidae) and their host plant availability in Costa Rica. Fla. Entomol. 71: 62-73. Jiron, L. F. and I. Hedstrom. 1991. Population fluctuations of the economic species of Anastrepha (Diptera: Tephritidae) related to mango fruiting phenology in Costa Rica. Fla. Entomol. 74: 98-105. Jiron, L. F. and R. G. Mexzon. 1989. Parasitoid hymenopterans of Costa Rica: Geographical distribution of the species associated with fruit flies (Diptera: Tephritidae). Entomophaga 34: 53-60. Kapatos, E., B. S. Fletcher, S. Pappas and Y. Laudeho. 1977. The release ofOpius concolor var. siculus (Hym.: Braconidae) against the spring generation ofDacus oleae (Dipt.: Tephritidae) on Corfu. Entomophaga 22: 265-270. Katiyar, K. P., J. M. Camacho, F. Geraud and R. Matheus. 1995. Parasitoides hymenopteros de moscas de las frutas (Diptera: Tephritidae) en la region occidental de Venezuela. Rev. Fac. Agron. (LUZ) (Maracaibo, Venezuela) 12: 303-312. Keiser, I., R. M. Kobayashi, D. H. Miyashita, E. J. Harris, E. L. Schneider and D. L Chambers. 1974. Suppression of Mediterranean fruit flies by oriental fruit flies in mixed infestations ofguava. J. Econ. Entomol. 67: 355-360. Knipling, E. F. 1992. Principles of insect parasitism analyzed from new perspectives: Practical implications for regulating insect populations by biological means. USDA Agric. Handbook 693, 337 pp.

PAGE 184

176 Lathrop, F. H. and R. C. Newton. 1933. The biology of Opius melleus Gahan, a parasite of the blueberry maggot. J. Agric. Res. 46: 143-160. Lawrence, P.O. 1981. Host vibrationA cue to host location by the parasite, Biosteres logicaudatus. Oecologia 48: 249-251. Lawrence, P. O., R. M. Baranowski and P. D. Greany. 1976. Effect of host age on development of Biosteres (=Opius) longicaudatus, a parasitoid of the Caribbean fruit fly, Anastrepha suspensa. Fla. Entomol. 59: 33-39. Lawrence, P. O., P. D. Greany, J. L. Nation and R. M. Baranowski. 1978. Oviposition behavior of Biosteres longicaudatus, a parasite of the Caribbean fruit fly, Anastrepha suspensa. Aim. Entomol. Soc. Am. 71: 253-256. Leonel Jr., F. L., R. A. Zucchi and R. A. Wharton. 1995. Distribution and tephritid hosts (Diptera) of braconid parasitoids (Hymenoptera) in Brazil. Int. J. Pest Management 41: 208-213. LeyvaVazquez, J. L. 1988. Temperatura umbra! y unidades calor requeridas por los estados inmaduros de Anastrepha ludens (Loew) (Diptera: Tephritidae). Folia Entomol. Mex. 74: 189-196. Liquido, N. J., E. J. Harris and L. A. Dekker. 1994. Ecology ofBactrocera latifrons (Diptera: Tephritidae) populations: Host plants, natural enemies, distribution, and abundance. Ann. Entomol. Soc. Am. 87: 71-84. Loni, A. 1997. Developmental rate of Opius concolor (Hym.: Braconidae) at various constant temperatures. Entomophaga 42: 359-366. Lopez, M., M. Aluja and J. Sivinski. Submitted. Hymenopterous larval-pupal and pupal parasitoids of Anastrepha species (Diptera: Tephritidae) in Mexican tropical deciduous and subdeciduous forests and agroecosystems. Biol. Control. Malavasi, A, and J. S. Morgante. 1981. Aduh and larval population fluctuation of Anastrepha fraterculus and its relationship to host availability. Environ. Entomol. 10: 275-278. Malo, E., P. S. Baker and J. Valenzuela. 1987. The abundance of species Anastrepha (Diptera: Tephritidae) in the coffee producing area of coastal Chiapas, southern Mexico. Folia Entomol. Mex. 73: 125-140. Meats, A. 1981. The bioclimatic potential of the Queensland fruit fly, Dacus tryoni, in Australia. Proc. Ecol. Soc. Aust. 11: 151-161.

PAGE 185

Messenger, P. S. and N. E. Flitters. 1954. Bioclimatic studies of three species of fixiit flies in Hawaii. J. Econ. Entomol. 47: 756-765. Messing, R. H. and E. B. Jang. 1992. Response of the fiiiit fly parasitoid D. longicctudata (Hymenoptera; Braconidae) to host-fruit stimuli. Environ. Entomol. 21; 1 189-1 195. Messing, R. H., J. M. Klungness and M. F. Purcell. 1994. Short-range dispersal of massreared Diachasmimorpha longicctudata and D. tryoni (Hymenoptera: Braconidae), parasitoids of tephritid fruit flies. J. Econ. Entomol. 87: 975-985. Messing, R. H., M. F. Purcell and J. M. Klungness. 1995. Short range dispersal of massreared Psyttalia fletcheri (Hymenoptera: Braconidae), parasitoids of Bactrocera cucurbitae (Diptera: Tephritidae). Environ. Entomol. 24: 1338-1343. Messing, R. H., L. M. Klungness, E. B. Jang, and K. A. Nishijima. 1996. Response of the melon fly parasitoid Psyttalia fletcheri (Hymenoptera: Braconidae) to host-habitat stimuli. J. Insect Behav. 9: 933-945. Monastero, S. and P. Delanoue. 1966. Lutte biologique experimentale contre la mouche de I'olive {Dacus oleae Gmel.) au moyen d'Opius concolor Szepl. siculus Mon. dans les lies Eoliennes (Sicile) en 1965. Entomophaga 11:41 1-432. Nasca, A. J. 1973. Parasitos de "Mosca de la Fruta" establecidos en algunas zonas de Tucuman. Rev. Agron. N. O. Argentina (Argentina) 10: 31-43. Nascimento, A. S., R. A. Zucchi, J. S. Morgante and A. Malavasi. 1982. Dinamica populacional das moscas-das-frutas do genero Anastrepha (Dip., Tephritidae) no Reconcavo Baiano 11 Fluctua9ao populacional. Pesq. Agropec. Bras. 17: 969-980. Newell, I. M. and F. H. Haramoto. 1968. Biotic factors influencing populations oi Dacus dorsalis in Hawaii. Proc. Hawaii. Entomol. Soc. 20: 81-139. Nguyen, R., C. Poucher and J. R. Brazzel. 1992. Seasonal occurrence of Anastrepha suspensa (Diptera: Tephritidae) in Indian River County, Florida, 1984-1987. J. Econ Entomol. 85: 813-820. Nishida, T. 1955. Natural enemies of the melon fly, Dacus curcurbitae Coq. in Hawaii. Ann. Entomol. Soc. Am. 48: 171-178. O'Loughlin, G. T., R A. East and A. Meats. 1984. Survival, development rates and generation times of the Queensland fruit fly, Dacus tryoni, in a marginally favorable cUmate: Experiments in Victoria. Aust. J. Zool. 32: 353-361. Ovruski, S. M. 1994. Comportamiento en la deteccion del huesped deAganaspis pelleranoi (Hymenoptera: Eucoilidae), parasitoide de larvas de Ceratitis capitata (Diptera: Tephritidae). Rev. Soc. Entomol. Argent. 53: 121-127.

PAGE 186

178 Ovruski, S. M. 1995. Pupal and larval-pupal parasitoids (Hymenoptera) obtained from Anastrepha spp. and Ceratitis capitata (Dipt.: Tephritidae) pupae collected in four localities of Tucuman Province, Argentina. Entomophaga 40: 367-370. Ovruski, S. M. and P. Fidalgo. 1994. Use of parasitoids (Hym.) in the control of fruit flies (Dip.: Tephritidae) in Argentina: A bibliographic review (1937-1991). lOBCAVPRS 17: 84-92. Palacio, I. P., A. G. Ibrahim and R. Ibrahim. 1991. Interspecific competition among opiine parasitoids of the oriental fruit fly, Bactrocera dorsalis (Hendel). Philipp. Entomol. 8: 1087-1097. Pemberton, C. E. and H. F. Willard. 1918. Interrelations of fruit fly parasites in Hawaii. J. Agr. Res. 12: 285-296. Petr, J. (ed ). 1991. Weather and yield. Developments in Crop Science 20. Elsevier, Amsterdam. 288 pp. Prescott, J. A. and R. M. Baranowski. 1971 . Effects of temperature on the immature stages Anastrepha suspensa (Diptera: Tephritidae). Fla. Entomol. 54: 297-303. Prokopy, R. J. and R. P. Webster. 1978. Oviposition-deterring pheromone ofRhagoletis pomonella. A kairomone for its parasitoid Opius lectus. J. Chem. Ecol. 4: 481-494. Purcell, M. F. 1998. Contribution of biological control to integrated pest management of tephritid fruit flies in the tropics and subtropics. Integ. Pest Manage. Rev. In Press. Purcell, M. F., C. G. Jackson, J. P. Long, and M. A. Batchelor. 1994. Influence of guava ripening on parasitism of the oriental fruit fly, Bactrocera dorsalis (Hendel) (Diptera: Tephritidae), Diachasmimorpha longicaudata (Ashmead) (Hymenoptera: Braconidae) and other parasitoids. Biol. Control 4: 396-403. Ramadan, M. M. and T. T. Y. Wong. 1990. Biological observations on Tetrastichus giffardianus (Hymenoptera: Eulophidae), a gregarious endoparasitoid of the Mediterranean fruit fly and the oriental fixiit fly (Diptera: Tephritidae). Proc. Hawaii. Entomol. Soc. 30: 59-62. Ramadan M. M., T. T. Y. Wong and J. W. Beardsley. 1992. Reproductive behavior of Biosteres arisarms (Sonan) (Hymenoptera: Braconidae), an egg-larval parasitoid of the oriental fruit fly. Biol. Control 2: 28-34. Ramadan, M. M., T. Y. Y. Wong, and J. C. Herr. 1994a. Is the oriental fruit fly (Diptera: Tephritidae) a natural host for the opiine parasitoid Diachasmimorpha tryorti (Hymenoptera: Braconidae)? Environ. Entomol. 23: 761-769.

PAGE 187

179 Ramadan M. M., T. T. Y. Wong and D. O. Mclnnis. 1994b. Reproductive biology of Biosteres arisanus (Sonan) (Hymenoptera: Braconidae), an egg-larval parasitoid of the oriental fruit fly. Biol. Control 4: 93-100. Ramadan M. M., T. T. Y. Wong and R. H. Messing. 1995. Reproductive biology of Biosteres vandenboschi (Hymenoptera: Braconidae), a parasitoid of early-instar oriental fruit fly. Biol. Control 2: 28-34. Raper, C. D. and P. J. Kramer (eds.). 1983. Crop reactions to water and temperature stresses in humid, temperate climates. 373 pp. Westview Press, Boulder, Colorado. Raspi, A. and A. Loni. 1994. Alcune note sull'allevamento di Opius concolor (Szepl.) (Hymenoptera: Braconidae) e su recenti tentativi d'introduzione della specie in Toscana ed inLiguria. Frustula Entomol. 30: 135-145. Roitberg, B. D. and R. G. Lalonde. 1991. Host marking enhances parasitism risk for a fruit-infesting fly Rhagoletis basiola. Oikos 61: 389-393. SAS Institute. 1982. SAS User's Guide: Statistics, 1982 Edition. 584 pp. SAS Institute, Cary, North Carolina. SAS Institute. 1989. SAS/STAT User's Guide, Version 6, Fourth Edition, Volume 2. 846 pp. SAS Institute, Cary, North Carolina. Sivinski, J. 1991. The influence of host fruit morphology on parasitization rates in the Caribbean fruit fly, Anastrepha suspensa. Entomophaga 36: 447-454. Sivinski, J. M. 1996. The past and potential of biological control of fruit flies, pp. 369375. In B. A. McPheron and G. J. Steck (eds ). Fruit fly pests: A world assessment of their biology and management. St. Lucie Press, Defray Beach, Florida. Sivinski, J. M., C. O. Calkins, R. Baranowski, D. Harris, J. Brambila, J. Diaz, R. E Bums, T. Holler and G. Dodson. 1996. Suppression of a Caribbean fruit fly {Anastrepha suspensa (Loew) Diptera: Tephritidae) population through augmented releases of the parasitoid Diachasmimorpha longicaudata (Ashmead) (Hymenoptera: Braconidae) Biol Control 6: 177-185. Sivinski, J., M. Aluja and M. Lopez. 1997. The spatial and temporal distributions of parasitoids of Mexican ^«a5/rep//a species (Diptera: Tephritidae) within canopies of fruit trees. Ann. Entomol. Soc. Am. 90: 604-618. Sivinski, J., M. Aluja, T. Holler and A. Eitam. 1998. A phenological comparison of two braconid parasitoids of the Caribbean fruit fly {Anastrepha suspensa (Loew)) (Diptera: Tephritidae). Environ. Entomol. In Press.

PAGE 188

180 Sivinski, J., T. Holler and F. Jeronimo. Submitted. Aerial releases of the Mediterranean fruit fly (Ceratitis ccpitata (Wiedemann)) parasitoids (Hymenoptera: Braconidae). Fla. Entomol. Snowball, G. J. 1966. Status of introduced parasites of Queensland fruit fly {Strumeta tryoni), 1962-1965. Aust. J. Agric. Res. 17: 719-739. Snowball, G. J. and R. G. Lukins. 1964. Status of introduced parasites of Queensland fruit fly {Strumeta tryoni), 1960-1962. Aust. J. Agric. Res. 15: 586-608. Snowball, G. J., F. Wilson, T. G Campbell and R. G. Lukins. 1962. The utilization of parasites of oriental fruit fly (Dacus dorsalis) against Queensland fruit fly {Strumeta tryoni). Aust. J. Agric. Res. 13: 443-460. Soto-Manitiu, J. and J. F. Jiron. 1989. Studies on the population dynamics of the fruit flies, Anastrepha (Diptera: Tephritidae) associated with mango {Mangifera indica L.) in Costa Rica. Trop. Pest Manage. 35: 425-427. Stark, J. D., R. I. Vargas and R. K. Thalman. 1991. Diversity and abundance of oriental fruit fly parasitoids (Hymenoptera: Braconidae) in guava orchards in Kauai, Hawaii. J. Econ. Entomol. 84: 1460-1467. Stark, J. D., R. I. Vargas and W. A. Walsh. 1994. Temporal synchrony and patterns in an exotic host-parasitoid community. Oecologia 100: 196-199. Stone, A. 1942. The fruit flies of the genus Anastrepha. USDA Misc. Publ. 439: 1-112. Tan, K. and M. Serit. 1994. Adult population dynamics of Bactrocera dorsalis (Diptera: Tephritidae) in relation to host phenology and weather in two villages of Penang Island, Malaysia. Environ. Entomol. 23: 267-275. Tauber, M. J., C. A. Tauber, J. R. Nechols and J. J. Obrycki. 1983. Seasonal activity of parasitoids: Control by external, internal and genetic factors, pp. 87-108. In V. K. Brown and I. Hodek (eds ). Diapause and life cycle strategies in insects. Dr W. Junk, The Hague. Tauber, M. J., C. A. Tauber and S. Masaki. 1986. Seasonal adaptations of insects. Oxford University Press, New York. Thomas, D B. 1997. Degree-day accumulations and seasonal duration of the preimaginal stages of the Mexican fruit fly (Diptera: Tephritidae). Fla. Entomol. 80: 71-79. Van den Bosch, R., H. A. Bess and F. H. Haramoto. 1951. Status of oriental fruit fly parasites in Hawaii. J. Econ. Entomol. 44: 753-759. Van den Bosch, R. and F. H. Haramoto. 1953. Opius oophilus Fullaway, an egg-larval parasite of the oriental fruit fly discovered in Hawaii. Proc. Hawaii. Entomol. Soc. 14:

PAGE 189

181 251-255. Van den Bosch, R. and F. H. Haramoto. 1951. Competition among parasites of the oriental fruit fly. Proc. Hawaii. Entomol. Soc. 15: 201-206. Van Dreische, R. G. and Bellows, T. S. Jr. 1996. Biological Control. Chapman and Hall, New York. 539 pp. Vargas, R. I., E. J. Harris and T. Nishida. 1983a. Distribution and seasonal occurrence of Ceratitis capitata (Wiedemann) (Diptera: Tephritidae) on the island of Kauai in the Hawaiian islands. Environ. Entomol. 12: 303-310. Vargas, R. 1., T. Nishida and J. W. Beardsley. 1983b. Distribution and abundance of Dacus dorsalis (Diptera: Tephritidae) in native and exotic forest areas on Kauai. Environ. Entomol. 12: 1185-1189. Vargas, R. 1., J. D. Stark and T. Nishida. 1989. Abundance, distribution, and dispersion indices of the oriental fruit fly and melon fly (Diptera: Tephritidae) on Kauai, Hawaiian islands. J. Econ. Entomol. 82: 1609-1615. Vargas, R. I., J. D. Stark and T. Nishida. 1990. Population dynamics, habitat preference, and seasonal distribution patterns of oriental fruit fly and melon fly (Diptera: Tephritidae) in an agricultural area. Environ. Entomol. 19: 1820-1828. Vargas, R. I., J. D. Stark, G. K. Uchida and M. Purcell. 1993. Opiine parasitoids (Hymenoptera: Braconidae) of oriental fruit fly (Diptera: Tephritidae) on Kauai island, Hawaii: Islandwide relative abundance and parasitism rates in wild and orchard guava habitats. Environ. Entomol. 22: 246-252. Vargas, R. I., W. A. Walsh and T. Nishida. 1995. Colonization of newly planted coffee fields: Dominance of Mediterranean fruit fly over oriental fhiit fly (Diptera: Tephritidae). J. Econ. Entomol. 88: 620-627. Walker, T. J., J. J. GaflTney, A. W. Kidder, and A. B. Ziffer. 1993. Florida Reach-Ins: Environmental chambers for entomological research. Am. Entomol. 39: 177-182. Weems, H. V. 1966. The Caribbean fruit fly in Florida. Fla. State Hortic. Soc 79 401405. Wharton, R. A. 1997. Generic relationships of opiine Braconidae (Hymenoptera) parasitic on fruit-infesting Tephritidae (Diptera). Cont. Am. Entomol. Inst. 30: 1-53. Wharton, R. A. and P. M. Marsh. 1978. New worid opiine (Hymenoptera: Braconidae) parasitic on Tephritidae (Diptera). J. Washington Acad. Sci. 68: 17-167.

PAGE 190

182 Wharton, R. A., F. E. Gilstrap, R. H. Rhode, M. Fischel-M and W. G. Hart. 1981. Hymenopterous egg-pupal and larval-pupal parasitoids of Ceratitis capitata and Anastrepha spp. (Diptera: Tephritidae) in Costa Rica fruit flies. Entomophaga 26: 285290. Willard, H. F. and A. C. Mason. 1937. Parasitization of the Mediterranean fruit fly in Hawaii, 1914-33. U.S. Dept. Agric. Cir. 439, 17 pp. Wong, T. T. Y. and M. M. Ramadan. 1992. Mass-rearing biology of larval parasitoids (Hymenoptera: Braconidae: Opiine) of tephritid fruit flies (Diptera: Tephritidae) in Hawaii, pp. 405-426. In T. E. Anderson and N. C. Leppa (eds.). Advances in insect rearing for research and pest management. Westview, Boulder, Colorado. Wong, T. T. Y. and M. M. Ramadan. 1987. Parasitization of the Mediterranean and oriental fruit flies (Diptera: Tephritidae) in the Kula area of Maui, Hawaii. J. Econ. Entomol. 80: 77-80. Wong, T. T. Y., N. Mochizuki and J. I. Nishimoto. 1984. Seasonal abundance of parasitoids of the Mediterranean and oriental fruit flies (Diptera: Tephritidae) in the Kula area of Maui, Hawaii. Environ. Entomol. 13: 140-145. Wong, T. T. Y., M. M. Ramadan, D. O. Mclnnis, N. Mochizuki, J. I. Nishimoto and J. C. Herr. 1991. Augmentative releases of Diachasmimorpha tryoni (Hymenoptera: Braconidae) to suppress a Mediterranean fruit fly (Diptera: Tephritidae) population in Kula, Maui, Hawaii. Biol. Control 1: 2-7. Zwdlfer, H. 1971. The structure and effect of parasite complexes attacking phytophagous host insects, pp. 405-418. In P. J. den Boer and G. R. Gradwell (eds ). Dynamics of populations: Proceedings of the Advanced Study Institute on "Dynamics of numbers in populations." Center for Agr. Pub. And Doc, Wageningen, The Netherlands.

PAGE 191

BIOGRAPHICAL SKETCH Avraham (Avi) Eitam was born in 1960 in Toronto, Ontario, Canada, and immigrated with his family to Israel in 1969. He graduated from Hugim High School in Haifa in 1978, and served in the Israel Defense Forces from 1979 to 1982. Upon completion of military service, he enrolled in undergraduate studies at Tel Aviv University, earning his B.Sc. in Biology Magna Cum Laude in 1985. He earned his M.Sc. in Zoology, also from Tel Aviv University, in 1989. The title of his thesis was "Aspects of social behavior in two species of halictine bees." Between 1989 and 1992, he worked in research and pest management at the Arava Research Station, Yotvata, Israel, concentrating on insect pests of the date palm. He enrolled in Ph.D. studies at the University of Florida in 1993. 183

PAGE 192

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and qiiality, as a dissertation for the degree of Doctor of Philosophy. jivinski. Chairman ^sistant Professor of Entomology and Nematology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarfy presentation and is fiilly adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Richard M. Baranowski Professor of Entomology and Nematology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Jonathan H. Crane Associate Professor of Horticultural Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. J. Howard Frank Professor of Entomology and Nematology

PAGE 193

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Patrick D. Greany Professor of Entomology and Nematology This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfilhnent of the requirements for the degree of Doctor of Philosophy. May, 1998 Dean, College?^ of Agriculture Dean, Graduate School