Butterfly Abundances, Larval Predation, and Egg Parasitism as Determined by Proximity to Butterfly Farms in Florida and ...

MISSING IMAGE

Material Information

Title:
Butterfly Abundances, Larval Predation, and Egg Parasitism as Determined by Proximity to Butterfly Farms in Florida and Costa Rica
Physical Description:
1 online resource (93 p.)
Language:
english
Creator:
Whelan, John Courtland
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Entomology and Nematology
Committee Chair:
Emmel, Thomas C.
Committee Co-Chair:
Daniels, Jaret
Committee Members:
Capinera, John L.
Miller, Jacqueline Y.

Subjects

Subjects / Keywords:
butterfly, costa, ecotourism, farms, florida, larvae, lepidoptera, oviposition, parasitism, predation, rica
Entomology and Nematology -- Dissertations, Academic -- UF
Genre:
Entomology and Nematology thesis, M.S.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
Butterfly farming is an emerging and rapidly growing industry, both in developed and developing nations around the world. As an industry focused on breeding butterflies for sale, while providing a number of incentives to local communities such as jobs, revenue, education, and conservation, these farms are becoming popular enterprises and ecotourism destinations. However, little is known about the impacts of these farms from a biological and ecological viewpoint. This study aimed to collect data on 1) wild butterfly abundance, as indicated by oviposition rates, 2) larval predation, and 3) egg parasitism rates to determine any effect that proximity to the farm epicenter might have on each factor. Additionally, the experiment was duplicated in north central Florida, U.S.A and Guanacaste, Costa Rica, to test for any similarities or differences in the aforementioned factors and variables that may be attributable to location. It was found that butterflies tended to oviposit in greater numbers closer to the farm epicenter in north central Florida, suggesting that the farm could serve as an attractive force for gravid female butterflies. Although untested, components of the farm, such as enhanced host and nectar plant biomasses and high levels of conspecific density, could contribute to the higher butterfly abundance closer to the farm. The effect these components have on butterfly abundances and the possibility of butterfly attraction to specific areas of the farm necessitates further research. Unexpectedly, there was no oviposition recorded during the entire Costa Rica trial. It is possible that visibility of the host plants and seasonal shifts in host plant preference by the gravid females severely affected oviposition in the tropical forest habitat. Larval predation was found to be more frequent in areas distal to the farm epicenters, both 100 meters from the farm as well as several kilometers. Lastly, in all trial sites, no egg parasitism was observed. These results suggest that while gravid female butterflies are more abundant closer to butterfly farms, there is less threat of predation to their offspring when ovipositing near the farm. The initial prediction that the farm may act as an ecological trap seems unlikely, as, in fact, larval survivorship is significantly higher near butterfly farms.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2008.
Local:
Adviser: Emmel, Thomas C.
Local:
Co-adviser: Daniels, Jaret.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31
Statement of Responsibility:
by John Courtland Whelan.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2008
System ID:
UFE0022690:00001


This item has the following downloads:


Full Text









































--_-~-
'' :L* .r .r


Figure 4-1. Although most host plants were situated in fields like this at Shady Oaks Butterfly
Farm, some plants were in more forested areas as seen in the background









conspecifie densities due to higher probabilities of locating suitable mates there. Specifically,

males would likely be attracted to areas of high female density, (Hanski et al. 1994, Kuussaari

1998). The second argument was given by Gilbert and Singer (1973), who argued that a high

density of butterflies in an area, especially of the same species, could be an indicator for suitable

habitat. Therefore, an area with high conspecifie density would trigger emigration from areas

with low conspecifie density as new adults would view the high density of same-species as a sign

of suitable habitat, perhaps containing ample resources such as water, nectar, and host plants.

Positively density-dependent dispersal

On the other hand, there are also arguments supporting positively density-dependent

dispersal, which is when individuals disperse at higher rates away from areas exhibiting high

conspecifie density (Dethier & MacArthur 1964; Odendaal et al. 1989; Baguette et al. 1996,

1998). These supporting arguments most often have to do with social interactions within the

patch (Shapiro 1970; Odendaal et al. 1985, 1989; Baguette et al. 1996, 1998). For instance, in

some species of butterflies, males will vigorously pursue females in order to mate with them.

While this may be effective for males that are pursuing un-mated females, sometimes the

females have already mated and are reluctant to mate again. This does not often dissuade the

males, as studies have observed them harassing females to the point of the female leaving the

patch (Odendaal et al. 1985). Should this happen frequently and repeatedly, a likely scenario in

a high density population, the result would be a high dispersal rate away from the area by

females. This, in turn, could trigger a similar dispersal pattern among males that may then leave

the area too, in pursuit of females or in search of an area with more females (Baguette et al.

1998).

A high conspecifie density in a particular patch may also directly affect habitat quality

of the patch, which could lead to higher emigration rates should the patch' s density surpass the









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

BUTTERFLY ABUNDANCES, LARVAL PREDATION, AND EGG PARASITISM AS
DETERMINED BY PROXIMITY TO BUTTERFLY FARMS
IN FLORIDA AND COSTA RICA

By

John Courtland Whelan

August 2008

Chair: Thomas C. Emmel
Cochair: Jaret C. Daniels
Maj or: Entomology and Nematology

Butterfly farming is an emerging and rapidly growing industry, both in developed and

developing nations around the world. As an industry focused on breeding butterflies for sale,

while providing a number of incentives to local communities such as jobs, revenue, education,

and conservation, these farms are becoming popular enterprises and ecotourism destinations.

However, little is known about the impacts of these farms from a biological and ecological

viewpoint. This study aimed to collect data on 1) wild butterfly abundance, as indicated by

oviposition rates, 2) larval predation, and 3) egg parasitism rates to determine any effect that

proximity to the farm epicenter might have on each factor. Additionally, the experiment was

duplicated in north central Florida, U.S.A and Guanacaste, Costa Rica, to test for any similarities

or differences in the aforementioned factors and variables that may be attributable to location. It

was found that butterflies tended to oviposit in greater numbers closer to the farm epicenter in

north central Florida, suggesting that the farm could serve as an attractive force for gravid female

butterflies. Although untested, components of the farm, such as enhanced host and nectar plant

biomasses and high levels of conspecific density, could contribute to the higher butterfly

abundance closer to the farm. The effect these components have on butterfly abundances and the








































Figure 4-6. Heliconia plants around the exterior of the farm epicenter could have acted as
competition for oviposition in the area










degradation. As stated previously, color of the host plant plays a maj or role in whether or not a

female butterfly is successful in finding an acceptable host (Prokopy & Owens 1983). As stated

by Myers (1985) and Wolfson (1980), color may depend on plant chemistry, which is directly

related to the physiological condition of the plant. It is possible that as the experiment

progressed, the physiological condition of the plants gradually rendered them less and less

visible, attractive, and suitable to gravid female butterflies.

Conspecific egg loads have been found to be oviposition deterrents for many species of

butterflies, including those in the genus Danaus (Rausher 1979b, Rothschild & Schoonhoven

1977). For this very reason, eggs were removed daily in an effort to minimize oviposition

deterrence. However, some species of Lepidoptera have been known to chemically detect the

presence of conspecifics (Renwick & Radke 1980). It is possible that residual chemicals from

the egg remained on the leaf surface, which may have acted as an oviposition deterrent.

Despite these possible explanations for a bell-shaped curve in oviposition for day and the

very significant p-value, we must not overlook the fact that this bell-shaped curve was observed

solely at Greathouse. Because Greathouse yielded significantly more data than Shady Oaks, the

analysis of the combined data overemphasized the results from Greathouse. However, there was

clearly a pronounced day effect at Shady Oaks as well, with fluctuations up and down throughout

the 10-day trial. Analysis of the day main effect suggests that more trials need to be conducted if

we are to confidently say that a bell-shaped pattern is indeed occurring.

Main Effect 2: Geographic Location

The effect that geographic location had on egg number was also very significant (p-value

< 0.0001). When the data are analyzed by location, we see that the Greathouse trial yielded

approximately three times more eggs overall than the Shady Oaks trial. Similarly, almost three









BUTTERFLY ABUNDANCES, LARVAL PREDATION, AND EGG PARASITISM AS
DETERMINED BY PROXIMITY TO BUTTERFLY FARMS
IN FLORIDA AND COSTA RICA




















By

JOHN COURTLAND WHELAN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2008










possibility of butterfly attraction to specific areas of the farm necessitates further research.

Unexpectedly, there was no oviposition recorded during the entire Costa Rica trial. It is possible

that visibility of the host plants and seasonal shifts in host plant preference by the gravid females

severely affected oviposition in the tropical forest habitat. Larval predation was found to be

more frequent in areas distal to the farm epicenters, both 100 meters from the farm as well as

several kilometers. Lastly, in all trial sites, no egg parasitism was observed. These results

suggest that while gravid female butterflies are more abundant closer to butterfly farms, there is

less threat of predation to their offspring when ovipositing near the farm. The initial prediction

that the farm may act as an ecological trap seems unlikely, as, in fact, larval survivorship is

significantly higher near butterfly farms.







































Figure 2-10. Control plot used in north central Florida experiments










increased awareness, which are key components in many commercial butterfly farming

operations.

These attractive qualities of butterfly farming render it very appealing to businessmen

and naturalists alike. However, an additional sector is now participating: rural communities in

developing nations. From a conservation perspective, this sector is very intriguing. As stated by

Omenge (2002), butterfly farming operations in developing nations are contributing to local

awareness and conservation of biodiversity, which is a very positive result. Although butterfly

farms are typically regarded as positive influences in the environment on a macro-scale, one

must not overlook the fact that they may influence their environment on a more local scale.

Because of the rapidly increasing numbers of farms and the close association that these farms

have with ecotourism, it is timely to assess some of the ecological processes, ecological impacts,

and general influences that they exert on their surrounding ecosystems.

A neoteric study by Gordon and Ayiemba (2003) incorporated the monitoring of butterfly

population abundances in the forest surrounding butterfly farms in Kenya in an effort to examine

the impact such farms have on the relative butterfly abundances in the area. After establishing a

highly successful butterfly farming operation aimed at deterring local communities from

destroying the nearby forest, the Kipepeo proj ect was born. Before the proj ect began, however,

butterfly abundance counts were taken. Then, after 4 years of farming operations (and after

$400,000 of farming revenue had come to the local community!, butterfly abundance counts

were taken again (Gordon and Ayiemba 2003). In both instances, these counts were taken by

researchers walking transects and counting sightings in each transect, which provided data on 60

species of butterflies. Of these 60 species, 23 were actively being reared at the farm at the time,

while 37 were not. Results showed that butterfly abundances were significantly higher after the










plants were in the field, they became degraded due to high levels of oviposition and stressful

environmental conditions such as intense sunlight, high temperatures, hard rainfall or wind

dehydration of leaves. A study by Ilse (1937) was the first to report that females in most families

of Lepidoptera engage in a "drmmming" action on the surface of the leaf on which they intend to

oviposit. Later described as a "scrapping" action by Fox (1966), Fox suggested that the spines

on the foretarsi abraded the leaf surface in order to release essential oils from the leaf, which

were likely used in host plant selection and acceptance. However, this supposed degradation of

the leaf surface is not limited to just female lepidopterans. Some male danaid butterflies have

been reported to scratch withered leaves in order to attain pyrrolizidine alkaloids from the plant

(Boppre 1983). Two studies, one by Schurr & Holdaway (1970) and the other by Renwick and

Radke (1982), suggest that injured and stressed plants may actually stimulate the release of

volatiles from host plants, which then act as oviposition deterrents to several species of

Lepidoptera. Should butterflies be instigating their host plant to release an oviposition deterrent

caused by plant damage and stress, we would expect high levels of oviposition to be followed by

a sharp decline, as seen in this experiment.

Environmental conditions may have also taken their toll on the plants, possibly rendering

them relatively unsuitable for oviposition. This experiment was conducted in August, which is

typically a very hot month in Florida, and it is not uncommon to see temperatures above 1000

Fahrenheit. Temperature readings during this experiment were consistent with historical records,

and temperatures in the high 90's were frequently observed. Summer storms are also common,

bringing strong wind gusts as well as flooding rain. Although these plants were cared for daily,

many were in direct sunlight for the maj ority of daylight hours and also exposed to the daily

pounding rain. It was evident that as time went on, some plants were showing signs of









plants (Rausher 1979a). For this reason, it is imperative to the survival of the species that the

female chooses an appropriate host plant correctly the vast maj ority of the time.

Habitat: Habitat is among the more important influences regarding host-plant selection

among insects. A habitat that simply contains specific host plants is not enough for that habitat

to be preferential to gravid female butterflies (Rausher 1979a). In fact, many ovipositing

butterflies do not use all of the habitats in which the appropriate host plants grow (Shapiro &

Carde 1970, Ehrlich et al. 1975). Egg laying patterns could be the result of patch size,

connectivity, and landscape matrix as well (Rabasa et al. 2005). Plant-insect interactions within

a site may also depend on processes occurring at larger spatial scales (Tscharntke & Brandl

2004). Nevertheless, patch size and landscape connectivity are postulated as key determinants

of butterfly presence in fragmented landscapes, due to the fact that smaller patches experience

higher rates of extinctions (James et al. 2003). A study by Rabasa et al. (2005) found that the

probability of egg presence and oviposition for a rare European butterfly did not differ

significantly between large and small patches.

Some previous studies suggest that oviposition search behavior continually evolves in

order to ensure that females lay eggs in the habitats that are most well suited for the growth and

development of the juvenile stages (Gilbert & Singer 1975, Wiklund 1977). As reported by

Hovanitz and Chang (1963), food plant choice appears to be at least partly genetically

determined in many insect species, with Lepidoptera being no exception. Therefore, one could

expect that differences in the suitability of food plants to influence the evolution of genes or

alleles controlling oviposition behavior (Rausher 1979a).

Forister (2004) tells us, however, that there is a positive correlation between female

preference of optimal host plant and larval performance and this has been found and cited in










the years, according to the farm manager. After 24 hours, the predator screens were removed.

Then 24 hours later, data were collected on the number of larvae remaining on each plant.

In addition, a control variable was used for this experiment at all geographic locations.

At a separate location away from the Butterfly farm, in similar habitat, a "false epicenter" was

marked with marking flags. This epicenter was the same size and dimensions as its experimental

counterpart, but did not have any butterfly farming operations taking place within its borders

(Figure 2-10). Two rows of host plants were set up in an identical fashion as described for the

experimental variable. The researcher oriented the rows so that both experimental and control

variables were along the same compass direction. Acclimation and data collection procedures

were identical to the procedure used for the experimental variable.









It was hypothesized that proximity to the farm would result in higher butterfly abundances so

that the highest abundances would be observed right next to the farm epicenter.

The second research goal was to investigate larval predation rates based on proximity to

the farm epicenter. It was hypothesized that larval predation rates would be highest at points

closest to the farm.

The third and final goal was to study the rate of egg parasitism, again based on proximity

to the farm epicenter. It was hypothesized that close proximity to the farm would yield higher

rates of egg parasitism.

Application of this study's results could serve to improve the ecological health of the

farm ecosystem, its surrounding area, and possibly result in the enhancement of overall

productivity. Butterfly farming operations, both domestically here in the U.S. as well as in

foreign countries, are often situated in relatively remote rural areas with significant amounts of

surrounding undeveloped land. Relative to urban or suburban settings, these farms typically

occupy a more pristine setting in which the surrounding ecosystem is healthier with higher levels

of biodiversity. Although this is not always the case, as some farms are indeed proximal to

urban areas, it is of great conservation concern to better understand the habits of all life stages of

butterflies living within proximity to butterfly farms in order to ensure that these enterprises

continue to contribute to biodiversity conservation and the principals of ecotourism.

Such a study is also beneficial to the butterfly farm as a way to increase understanding of

how butterflies are parasitized, preyed upon and their oviposition habits. By quantifying

parasitism rates, the farming operation may be able to adjust management practices in order to

ensure higher productivity through lowering parasitism rates of the species actively being bred.

This higher productivity, at a human social and community level, could result in more revenue









ACKNOWLEDGMENTS

First, I would like to thank my Mom, who has always been there to offer good advice,

encouragement and support. I am forever grateful for her guidance and dedication throughout

my life.

I would also like to thank Frank and Kim Powell for greatly contributing to my interest in

science and ecotourism. The opportunities they gave me provided a spark that quickly ignited

into a great passion.

My entire committee has been of so much help during my graduate education, and I cannot

thank them enough for providing me the opportunity to learn from each of them. Thanks go to

all those who have advised me along the way, including Tom Emmel, Jaret Daniels, Jackie

Miller, John Capinera, and Taylor Stein.

Edith and Steven Smith of Shady Oaks, Zane Greathouse of Greathouse Butterfly Farm,

and Ernesto Rodriguez and John Fazzini of El Bosque Nuevo were an essential part of my

research, as they provided me with help, guidance, and resources that were critical to the success

of my study. Their participation, cooperation, and knowledge is greatly appreciated.

Lastly, I would like to thank the Chair of my committee and maj or advisor, Tom Emmel,

who has provided me with incredible opportunities that have greatly contributed to my education

and have given me tremendous experiences in the world of ecotourism entomology. He is a

great mentor and friend, and a most sincere thank you is certainly in order.










comparing the experimental plot with the control plot, as well as when comparing the near

variable with the far variable at the experimental plot.

Higher levels of predation on larvae that are situated further from the butterfly farm could

be the result of either a higher presence of predators or higher rates of larval detection by

predators. It could also be possible that there are multiple populations of predators surrounding

the farm, with separate populations predating upon the larvae at different distances from the farm

epicenter. Perhaps the predator populations situated closer to the farm predate mainly on larvae

in and around the farm's gardens, nurseries, etc., and were satiated without needing to predate

upon the larvae presented to them in this experiment. In addition, it is possible the predator

population away from the farm, not typically having easy prey such as the larvae that were

placed in this experiment, took advantage of the opportunity to predate upon these easy prey

items. It could have also been possible that different species of predators were at the different

distances from the farm. Perhaps high predation rates at the control plot were due to one species,

while predation at the farm plot was from another species. The control plot in Florida, although

similar habitat to the farm plot, was unmowed, brushy and in very sandy soil. Solenopsis ants

were observed here and could have been the main predator involved. Similarly, predation at the

farm could have differed depending on the species of potential predator as well. Although

Solenopsis ants were also found here, they could have been more contained due to the mowing

regime (approximately once a week), rendering them a subordinate predator.

Lastly, relatively high levels of human activity could have affected predation levels. It is

possible that the near constant presence of ecotourists and farm workers around the farm grounds

was enough to impose a "threat" to avian and lizard predators (Figure 4-5). By trying to avoid

humans, potential predators would be indirectly deterred away from the farm. Should this be









CHAPTER 6
CONCLUSIONS

The results from this experiment show that there is indeed an effect that the typical

butterfly farm has on wild butterfly oviposition, and in turn, wild butterfly abundances. The

tendency for butterflies to be in greater abundances closer to the farm could be the result of a

combination of limited dispersal among butterflies from within the farm, whether escaped or

hatched in the nurseries, as well as wild butterflies being attracted to the farm epicenter and the

butterfly resources it contains. It is possible that the farm is emitting specific cues indicating

habitat quality as well as the presence of viable mates, although untested in this experiment.

This experiment cannot prove or disprove that these cues are real, nor if there is indeed a force

attracting wild butterflies. However, we can say that there are greater butterfly abundances

closer to the farm, whether from wild populations or butterflies from within the farm. It would

be a very interesting and helpful follow-up study to examine these cues and the attractive stimuli

these farms may elicit with regards to wild butterflies.

By testing for larval predation around the farm, we see that the farm poses no clear threat

of abnormally high levels of predation closer to the farm. In fact, the opposite is the case. There

appears to be abnormally low levels of predation as one gets closer to the farm. This could be

due to discrepatncies in predator populations and searching behavior. It could also be the result

of heightened levels of human activity, in the form of farm workers and ecotourists, which may be

perceived as a threat by potential predators.

By testing for egg parasitism, the data suggest that proximity to the farm does not elicit

higher levels of parasitism when compared to more distal areas. However, zero parasitism

compared with zero parasitism does not provide convincing enough statistics to authoritatively

claim no difference.









Finally, the location effect is shown below in Figure 3-6. There was over a 3-fold

increase in total oviposition from Shady Oaks to Greathouse, with 288 and 997 eggs deposited,

respectively, over the 10-day period.

Main Experiment: Butterfly Oviposition in Costa Rica

Oviposition in the Costa Rica trial was very surprising, as absolutely no eggs were

oviposited on any of the hostplants during the entire 10-day duration of the experiment.

Although this was an unexpected result, it does raise some interesting questions, which will be

addressed in the Discussion chapter.

Larval Experiment: Larval Survivorship as Determined by Distance from Farm Epicenters
in North Central Florida

The larval predation experiment suggests that there is indeed an effect that the butterfly

farm has on its environment, but perhaps not in the same way as was originally predicted.

After running the experiment at the farm and running a control in a field of similar habitat

approximately 5 miles away from the farm, several comparisons were made. First, total larval

survivorship (i.e., not accounting for distance from the epicenter) at the experimental plot was

compared to a control plot. Higher levels of survivorship among larvae at the experimental plot

were found when compared to the control plots (Figure 3-7). Survivorship at the experimental

plot was 41%, while survivorship at the control plot was 14% (Table 3-3). Using a T-test at

alpha = 0.005, the researcher was able to conclude that there is a significant difference between

the experimental and control plots, in terms of larval survivorship. Second, it was determined

that there was also a significantly higher level of survivorship for larvae situated closer (i.e.,

located at transect station 1, which was adj acent to the perimeter of the farm epicenter) to the

farm epicenter, when compared to larvae situated farther (i.e., located at transect station 11,

which was 100 meters away from the perimeter of the farm epicenter) from the farm epicenter











AND 1975. Butterfly ecology. Annual Review of Ecology and
Systematics. 6: 465-397.

GINGRAS, D., AND G. BolvlN. 2002. Effect of plant structure, host density and foraging duration
on host finding by Trichogramnma evanescens (Hymoptera: Trichogrammatidae).
Environmental Entomology 31: 1153-1157.

GLENDINNING, J. I., AND L. P. BROWER. 1990. Feeding and breeding responses of five mice
species to overwintering aggregations of the monarch butterfly. Journal of Animal
Ecology 59: 1091-1112.

GORDON, I., AND W. AYIE1VBA. 2003. Harnessing butterfly biodiversity for improving
livelihoods and forest conservation: The Kipepeo Project. Journal of Environment &
Development 12 (1): 82-98.

GOTTHARD, K. 1999. Life history analysis of growth strategies in temperate butterflies. Ph.D.
Thesis, Stockholm University.

.2000. Increased risk of predation as a cost of high growth rate: an experimental test
in a butterfly. Journal of Animal Ecology 69: 896-902.

HANSKI, I. AND M. E. GILPIN. 1997. Metapopulation Biology: Ecology, Genetics and Evolution,
Academic Press, San Diego.

.1999. Metapopulation Ecology, Oxford University Press, Oxford, UK.

HARRIS, M. O., M. SANDANYAKA, AND W. GRIFFIN. 2001. Oviposition preferences of the Hessian
fly and their consequences for the survival and reproductive potential of offspring.
Ecological Entomology 26: 473-486.

HOVANITZ, W., AND V. C. S. CHANG. 1 963. Change of food-plant preference by larvae of Pieris
urapae controlled by strain selection, and the inheritance of this trait. Journal of
Research on the Lepidoptera 1: 163-168.

ILSE, D. 1937. New observation on responses to colors in egg-laying butterflies. Nature 140:
544-545.

JA1VIES, M., F. GILBERT, AND S. ZALAT. 2003. Thyme and isolation for the Sinai baton blue
butterfly (Pseudophilotes sinaicus). Oecologia 134: 445-453.

JANZEN, D. H. 1967. Synchronization of sexual reproduction of trees with the dry season in
Central America. Evolution 21: 620-637.

KLO1VP, H., B. J. TEERINK. 1962. Host Selection and Number of Eggs per Oviposition in the Egg-
Parasite Trichogramma embryophagum. Nature 195: 1020-1021.









on the possible negative effects of lepidopteran parasites, particularly those parasitoids affecting

butterfly farming, captive breeding, and butterfly ecotourism programs.

When dealing with parasitoids that affect insects in the order Lepidoptera, butterfly

farmers and biologists alike can identify several types of insects that could pose potential

problems, such as Tachinid flies (Diptera: Tachinidae) and Braconid wasps (Hymenoptera:

Bradonidae), there is one key wasp group that overwhelmingly presents the greatest threat:

species of the egg parasitoid, genus Trichogra~nmna (Hymenoptera: Trichogrammatidae). Many

species in this genus are important natural enemies and are used in inundative and inoculative

biocontrol programs throughout the world (Li 1994, Smith 1996). While many species in the

genus Trichogra~nmna are generalists, the better-known species are specialists on parasitizing

eggs of the members of the order Lepidoptera (Pinto & Stouthamer 1994). Because of their

extremely small size, ranging from 0.2 to 1.5mm, these tiny wasps present a much greater

problem than other parasitoids for the captive breeding of Lepidoptera (Pinto & Stouthamer

1994). This is because exclusionary tactics such as screens, netting, and cages need to be almost

perfect (i.e., no gaps, tears, etc.) to ensure that these nearly microscopic insects do not have

access to breeding centers and the Lepidoptera eggs within. While newer exclusionary

technology allows for better parasite control, the costs of these more effective methods are often

prohibitive to butterfly farmers in developing nations (Gordon & Ayiemba 2003). Because

butterfly farming and ecotourism in developing nations are increasingly becoming vital sources

of revenue and a means for biodiversity conservation, a better understanding of egg parasitoid

behavior in these areas is critical.

Parasitoid Behavior

The behavior of Trichogra~nmna parasitoids was studied as early as 1935 by G. Salt and J.

Laing. Extensive experiments and reviews have been published subsequently, which have










number of eggs deposited at transects 9, 10, and 11 at Shady Oaks, the full 12 replications were

not able to be conducted. Rather, the sample size was 6, 9, and 6, respectively.


































Figure 2-8. M~orpho peleides host plant Leguminosa cristobal before and after predator screen
application


Figure 2-9. Close-up view of predator screen









that eggs deposited for 72 hr elicited and even more significantly stronger response and

suggested that substances associated with egg deposition could have diffused into the leaf tissue

or the leaf s wax layer, triggering a chemical response by the leaf. This chemical, in turn, would

at very least arrest, and possibly attract parasitoids. In my experiment, eggs were left on the

plant for no more than 24 hr, which may not have been enough time for synergistic effects

between the plant and the egg to attract parasitoids. The reason that eggs were removed daily

was to prevent oviposition competition in the oviposition study. Research has suggested that

gravid females may pass up suitable host plants due to the presence of conspecific eggs (Rausher

1979b, Rothschild 1977).

The second problem is that the plants were spaced, from a parasitoid's perspective,

relatively far apart from each other. This prevented parasitoids from dispersing by means of

walking and jumping, which is a vital means of locomotion for butterfly parasitoids (Pak et al.

1985). While nurseries inside the farm epicenter provided for high connectivity between host

plants, the 10-meter spacing used in the experiment provided relatively no connectivity. Gmngras

and Boivin (2002) stated that connectivity between host plants and host plant parts best explains

variability in parasitism rates.

Due to these reasons, it is likely that parasitoid presence is concentrated to areas within

the farm epicenter. Areas in which eggs are left on host plants for greater than 48 hrs, as well as

high levels of plant connectivity and homogeneity, are likely to present ideal habitats for

populations of egg parasitoids, thus, there was little need for them to disperse to the outer areas

of the farm (Lukianchus & Smith 1997). It would be beneficial to conduct further studies from

within the farm epicenter and select specific habitats and specific groups of plants rather than the

entire farm, as was done in this experiment.










females in the vicinity just as other plants were that were not located near this Monarch butterfly

enclosure.

Main Effect 3: Transect Station

The effect of transect station on oviposition was not as statistically significant as either

day or location, but still received a relatively low p-value (p-value = 0.2049) and interesting

summary statistics.

From the data, we found that there is a positive correlation between high oviposition

levels and proximity to butterfly farm epicenters. With oviposition also serving as an indicator

for butterfly abundance, the question must be posed as to why butterflies would tend to be more

abundant within 40, perhaps 30, meters from the perimeter of the farm epicenter?

As stated by Rausher (1979a), females that oviposit on plants yielding low larval

survivorship or poor larval growth will likely leave fewer descendents than females that oviposit

on the more appropriate host plants. For this reason, the choice that a female butterfly makes on

where to oviposit is critical. Although exceptions have been found, it is widely accepted that

oviposition search behavior has evolved to ensure that females lay eggs in the habitats that are

most suitable for the growth and development of the juvenile stages (Gilbert & Singer 1975,

Wiklund 1977). Therefore, gravid females may prefer the plants closer to the farm epicenter

because these plants are perceived by the female to be more suitable for development of their

offspring. Evidently, as the butterflies evaluated the balance between positive and negative

signals from the plant and external stimuli, such as presence of eggs and other insects, the plants

closer to the farm epicenter are more suitable. Although unproven here, it is possible that the

farm may be acting as a magnet, both attracting butterflies from wild populations, as well as

retaining any farm-raised or escapee butterflies from within the farm itself.









Once the experiment was set up, data were then collected on the number of new eggs

found on each plant per day (i.e., every 24 hours). These eggs were subsequently used in the

next part of the experiment (see Egg Parasitism section Page 37). All eggs were removed at

the time of data collection each day. This was to prevent gravid females from passing up these

plants when ovipositing in the following days, due to the possibility of females of that species

having responded to past selection to relieve conspecific egg competition by visually inspecting

host plants for the presence of previously laid eggs. Data collection was repeated on a daily

basis (i.e., every 24 hours) over a 10-day period resulting in 30 replicate host plant at each

transect number (distance from epicenter) for each farm location.

Egg Parasitism

The second part of the experiment was to test for egg parasitism at varying distances

from the butterfly farm epicenter. Using eggs from the previous oviposition experiment, a total

of 12 eggs was collected at random from each transect number and retained for use in this egg-

parasitism study. Eggs were collected from the plants every 24 hours, at 8:00 A.M. (this samples

eggs exposed to parasitism incidence from the preceding 24 hours). After collecting these eggs,

they were placed individually in plastic containers covered with a fine mesh rubber-banded on

top to allow air flow but to prevent any exit of the hatched larvae, or hatched parasite. The mesh

also blocked entrance of subsequent searching parasites. Then, the eggs were observed on a

daily basis until emergence of larvae (roughly four days), or parasitoid (Figure 2-5). Data were

then collected on whether butterfly larvae or a parasite emerged.

Larval Predation

The final part of the experiment was to test predation effects on larvae of Monarch

butterflies (Danaus plexippus) in north central Florida and Common Morpho butterflies (M~orpho

peleides) in Costa Rica. For each location, two host plant rows were placed radiating out from










Previous research suggests that when female butterflies are confronted with an array of potential

hosts, they exhibit a hierarchy in preferences, laying most eggs on the preferred plant, and fewer

on the next, and so on (Courtney et al. 1989, Thompson 1988).

A second scenario possibly explaining the difference between station groups 1-4 and 5-11

could be that egg number data (i.e., oviposition) may have come from butterflies from within the

farm rather than wild butterflies. These "within farm" butterflies could be ones that have

escaped from the enclosures, or possibly hatched from eggs that remained undetected by farm

workers as the host plants are moved around the farm. Oviposition from escapee and other farm-

raised butterflies may be more contained to the interior of the farm epicenter, depending on

location of the particular rearing cage from which the butterfly escaped or the host plant from

which the butterfly emerged. Previous studies have found that visual conspicuousness of host

plants is related to increased oviposition (Courtney 1982). Therefore, plants located at closer

transects may be more conspicuous to the escapee and farm-raised butterflies, and would likely

exhibit higher levels of oviposition. Although this may apply to the steady decline in egg

number from transect stations 1 through 7, it does little to explain the drastic drop from transect

station 4 to transect station 5.

The conspecific attraction hypothesis states that high-density populations yield higher

immigration rates to, and lower emigration rates away from, the habitat in question (Ray et al.

1991). Should these escapee butterflies be content with the available resources of the farm,

many may reside inside the farm epicenter. The likelihood of wild butterflies being present in

the outdoor nurseries and butterfly gardens, in addition to these escapee butterflies, could serve

as motivation for butterflies to remain in the area, according to the conspecific attraction

hypothesis.




































Figure 3-4. Location and day affected egg number more than just the sum of their individual
roles

300


250


200


150


100


50


1 2 3 4 5 6 7 8 9 10
HTotal Egg Count 126 110 158 201 189 252 104 58 58 29
Day

Figure 3-5. The day effect on egg number was relatively bell-shaped, with a peak in egg counts
for days 4 through 6 after initial placement of the potted plants









studies of various insect families (Mayhew 1997, Harris et al. 2001, Craig & Ohgushi 2002).

Nevertheless, this presumption of a strongly positive correlation has been integrated in many

biological models and taken as fact in many instances. Also termed the naive adaptationist

hypothesis, by Courtney and Kibota (1990), this is the most common hypothesis in regards to the

selection of host plants by phytophagous insects. It states that natural selection effectively

promotes female preference such that eggs are laid on the host plant that is the best host for her

offspring (Levins & MacArthur 1969). Reiterating what Raucher (1979a) said, there are two

ways that the plant can serve as best host: one way resulting in the highest survivorship rate

among offspring, and the other being the relative size of the offspring without regard to overall

survivorship rates (Rausher 1979a).

Oviposition is used in this study as an indicator for butterfly abundance. By

standardizing the oviposition substrate (i.e., host plant), and environmental conditions such as

seasonality and landscape connectivity, any fluctuations in the number of eggs deposited should

indicate a fluctuation in butterfly abundance in the area.












4 DI SCUS SSION ................. ...............59................


Butterfly Abundance and Oviposition at North Central Florida Farms ................ ...............59
Main Effect 1: Day ............... .... ...............62.
Main Effect 2: Geographic Location ............ .....___ ...............64.
Main Effect 3: Transect Station ............_ ..... ..__ ...............67.
Larval Predation in Florida and Costa Rica............... ...............68..

Egg Parasitism in Florida and Costa Rica .............. ..... ....... ................7
Butterfly Abundance and Oviposition at the Northern Costa Rica Farm ............... ..............72

5 CONCLUSIONS .............. ...............83....


LIST OF REFERENCES ........._...... ...............85._.._. ......


BIOGRAPHICAL SKETCH .............. ...............93....




































To my Mom, who has been a great role model and friend









their corresponding F-values and P-values, all at preset alpha = 0.100. Transect effect was found

to have a p-value of 0.2049, which is slightly larger than the pre-set alpha level. However, this is

still a relatively low value. In conjunction with the summary statistics described previously,

there is enough evidence to suspect that the transect number (i.e., distance from the epicenter)

has an effect on butterfly oviposition.

Due to the strong effect that location and day had on egg number, it merited further

examination of the data based on individual locations and individual days. See Figure 3-2 and

Figure 3-3 for separate graphs for average egg number per day at each location in north central

Florida. From the data shown in Figure 3-2 and Figure 3-3, we can see a definite difference

between locations. Shady Oaks has a more pronounced negative correlation between egg

number and transect, while for Greathouse, egg number drops sharply around transect 5 and 6,

but then rises from 6 to 11 (Figure 3-2 and Figure 3-3).

As mentioned before, there is an interaction effect between location and day. Figure 3-4

shows the egg number broken down by individual days as well as the individual farm. This

interaction means that the combination of day and location yields more variance in egg number

than simply the sum of their individual main effects. We see that Shady Oaks yielded overall

fewer numbers of eggs in a vacillating fashion, while Greathouse exhibited a bell-shaped pattern,

with most oviposition occurring on day 5 and 6, with a sharp drop off on day 7 through 10.

The day effect was quite pronounced as well. Most eggs were deposited during the

middle of the experiment, with days 4, 5 and 6 seeing the most eggs at 201, 189 and 252 eggs

respectively (Figure 3-5). Then, on day 7, a sharp decline in oviposition was observed, which

then continued until the end of the experiment.










SCHURR, K., AND F. G. HOLDAWAY. 1970. Olfactory responses of female Ostrinia nubilalis
(Lepidoptera: Pyraustinae). Entomologia Experimentalis et Applicata 13: 455-461.

SHAPIRO, A. M. 1970. The role of sexual behavior in density-related disperal of pierid
butterflies. American Naturalist 104: 367-372.

,AND R. T. CARDE. 1970. Habitat selection and competition among sibling species of
satyrid butterflies. Evolution 24: 48-54.

SHREEVE, T. G. 1995. Butterfly mobility. Dr A. S. Pullin (Ed.). Ecology and conservation of
butterflies, pp. 37-45. Chapman and Hall, London.

SINGER, M. C. 1993. Behavioral constrains on the evolutionary expansion of an insect diet: a
case history. In L. Real (Ed.) Behavioral Mecha~nisms in Evohitionazy Ecology
Chicago: University of Chicago press.

SMInTH, C. C., AND FRETWELL. 1974. The optimal balance between size and number of offspring.
American Naturalist 108: 499-506.

SMInTH, S. M. 1996. Biological control with Drichogra~nmna: advances, successes, and potential of
their use. Annual Review of Entomology 41: 375-406.

STANTON, M. L. 1984. Short-term learning and the searching accuracy of egg-laying butterflies.
Animal Behavior 32: 33-40.

STEARNS, S. C. 1992. The Evohition of Life Histories, 1st edition, Oxford University Press,
Oxford, UK.

STEINGENGA M. J., AND K. FISCHER. 2007. Ovarian dynamics, egg size, and egg number in
relation to temperature and mating status in a butterfly. Entomologia Experimentalis et
Applicata 125: 195-203.

SURVEKROPP, B. P. 1997. Host-finding behavior of Trichogra~nmna brassicae in maize. Ph.D.
Dissertation, Wageningen Agricultural University, Wageningen, The Netherlands.

THOMSON, J. N. 1988. Variation in preference and specificity in monophagous and oligophagous
swallowtail butterflies. Evolution 42: 118-128.

TSCHARNTKE, T., AND R. BRANDAL. 2004. Plant-insect interactions in fragmented landscapes.
Annual Review of Entomology 49: 405-430.

WAJNBERG, E. 1994. Intra-population genetic variation in Drichogra~nmna. Dr E. Wajnberg, and
S. A. Hassan (Eds.). Biological control with egg parasitoids, pp. 245-271. Berkshire:
CAB International.









CHAPTER 1
INTTRODUCTION

Butterfly Farming and Ecotourism

The breeding of butterflies for sale, known as butterfly farming, is increasingly regarded

as a profitable microenterprise business. This is partly due to the fact that the overall use and the

diversity of applications for these alluring lepidopterans are expanding. Today, butterflies are

utilized in many ways, ranging from live butterfly releases at weddings to being displayed in live

butterfly exhibits in countries around the world (Young 1986). In addition to this increasing

popularity of these insects, techniques for farming, breeding, and rearing Lepidoptera are

becoming more fine-tuned, therefore expanding output and ultimately escalating the success of

the individual farmer. Organizations such as the International Butterfly Breeders Association

(IBBA), founded in 1998, are contributing vast amounts of information to the overall scene. The

IBBA coordinates workshops, manuals, and tutorials, allowing amateurs with little to no prior

knowledge to initiate breeding operations and rapidly become proficient at farming butterflies.

The resulting accessibility to butterfly farming through a virtually endless supply of knowledge

and support, combined with a relatively high demand for these insects, make this industry

increasingly more profitable and attractive. With some individual farms seeing annual revenues

in the 6-digit range, it is easy to see it becomes desirable to enter this multi-million dollar

industry.

At the same time, butterfly farming presents a novel situation in which to promote

ecotourism. Defined as "responsible travel to natural areas in which to conserve the environment

and promote the welfare of local people," ecotourism is intrinsically associated with

conservation (TIES 1998). Ecotourism is a powerful tool to promote conservation, both directly

through investment in conservation programs as well as indirectly through education and










LIST OF FIGURES


FiMr page

2-1 Entrances to the three butterfly farms used throughout the experiment ................... .........39

2-2 Diagram of farm epicenter (green circle) with radiation rows of hostplants at 10
meter transects (yellow dots) .............. ...............40....

2-3 Dense jungle causing one row of host plants to be shifted slightly ............. ........._.....41

2-4 Several plants needed to be placed close to roads in the Florida experiments ................42

2-5 Screened vials and stereoscope used to observe the emergence of parasites or larvae .....43

2-6 Diagram of larval predation experiment ......._..__ .. ..... ..._ ......_._..........4

2-7 Controlled environment used to acclimate Monarch larvae to their host plant .................44

2-8 M~orpho peleides host plant Leguminosa cristobal before and after predator screen
application............... ..............4

2-9 Close-up view of predator screen ......._.........._.. ......._. .........__.........45

3-1 Summary graph of average egg number for each transect............... ...............53

3-3 Mean egg number at each transect for Shady Oaks Butterfly Farm .........._...._ ...............54

3-4 Location and day affected egg number more than just the sum of their individual
roles............... ...............55.

3-5 The day effect on egg number was relatively bell-shaped, with a peak in egg counts
for days 4 through 6 after initial placement of the potted plants .............._ ................. 55

3-6 Location had a pronounced effect on egg number, with almost 3 times as many eggs
at Greathouse when compared to Shady Oaks ................. ...............56........... ..

3-7 Mean % survivor ship among larvae at experimental and control plots for north
central Florida farms ................ ...............57........... ....

3-8 Mean % survivor ship among larvae at far and near distances in the experimental plot
for north central Florida farms ................. ...............57........... ...

3-9 Mean % survivor ship among larvae at far and near distances in the control plot for
the northern Costa Rica farm .............. ...............58....

4-1 Although most host plants were situated in fields like this at Shady Oaks Butterfly
Farm, some plants were in more forested areas as seen in the background.......................76










WERNER, E. E., AND B. R. ANHOLT. 1993. Ecological consequences of the trade-off between
growth and mortality rates mediated by foraging activity. American Naturalist 142:
242-272.

WIKLUND, C. 1977. Oviposition, feeding and spatial separation of breeding and foraging
habitats in a population of Leptidea sinapis (Lepidoptera). Oikos 28: 56-68.

WILLIAMS, B. K., AND J. D. NICHOLS. 1984. Optimal timing in biological processes. American
Naturalist 123: 1-19.

WOLFSON, J. L. 1980. Oviposition response of Pieris urapae to environmentally induced variation
in Bra~ssica nigra. Entomologia Experimentalis et Applicata Entomologia
Experimentalis et Applicata 27: 223-232.

YAMPOLSKI L. Y., AND S. M. SCHEINER. 1996. Why larger offspring at lower temperatures? A
demographic approach. American Naturalist 147: 86-100.

ZALUCKI, M. P., AND R. L. KITCHING. 1982. Temporal and spatial variation of mortality in field
populations of Danaus plexippus L. and D. chrysippus L. larvae (Lepidoptera:
Nymphalidae). Oecologia 3: 201-207.

,S. B. M1ALCOLM, T. D. PAINE, C. C. HANLON, L. P. BROWER, AND A. R. CLARKE. 2001.
It' s the first bites that count: Survival of first-instar monarchs on milkweeds. Austral
Ecology 26 (5): 547-555.









times more eggs were found, on average, for each transect. For certain days, almost 10 times

more eggs were found at Greathouse versus Shady Oaks.

One possible reason for this difference is the farm epicenter size, which was observed by

the researcher to be directly proportional to the number of nurseries, butterfly rearing houses, and

other parts of the farm involved in butterfly production. The farm epicenter for Shady Oaks was

nearly 7,000 square meters, while that of Greathouse was just over 12,600 square meters. With

twice the total area, occupied by nectar plants, host plants, and butterfly enclosures at

Greathouse, we see that there may be a positive correlation between the size of the butterfly farm

and rates of butterfly abundances around the farm. The greater size and number of butterfly

enclosures could increase the potential for butterflies escaping on the principle that more

enclosures create more risk of escape.

A second difference between the two farms was the landscape surrounding the farm

epicenter, and there were two maj or differences. First is the amount of forested land surrounding

the farm. The type of forested land was similar at both farms, consisting of mostly pine and oak

trees, with some undergrowth, all growing in sandy, well-drained soil. Shady Oaks was bordered

by far more forested areas with all but two rows traversing part of the surrounding forest.

Although the maj ority of host plants were placed in the grassy areas that occupied most of the

experimental plot, some plants were placed in the forested land, rendering them much less visible

to gravid female butterflies (Figure 4-1). Conversely, Greathouse butterfly farm was situated in

a way that only one row of hostplants traversed forested land. The vast maj ority of plants were

situated in well-kept grass, allowing the visual cues to be better perceived by searching

butterflies (Figure 4-2). The second landscape characteristic that may have led to greater

oviposition rates at Greathouse was the presence of a residence within close proximity to the










KUUSSAARI, M., M. NIEMINEN, AND I. HANSKI. 1996. An experimental study of migration in the
Glanville fritillary butterfly M~elitaea cinxia. Journal of Animal Ecology 65: 791-801.

,M., I. SACHERI, AND M. CAMVARA. 1998. Allee effect and population dynamics in
the Glanville fritillary butterfly Oikos 82: 384-392.

KYI, A., M. P. ZALUCKI, AND I. J. TITMARSH. 1991. An experimental study of early stage survival
of Helicoverpa armigera (Lepidoptera: Noctuidae) on cotton. Bulletin of
Entomological Research 81: 263-271.

LEVINS, R. 1968. Evolution in Changing Environments. Princeton University Press, New Jersey.

,AND R. H. MACARTHUR. 1969. An hypothesis to explain the incident of monophagy.
Ecology 50: 910-911.

LI, Y. L. 1994. Worldwide use of Trichogramnma for biological control of different crops: a
survey. In E. Wajnberg, and S. A. Hassan (Eds.). Biological control with egg
parasitoids, pp. 37-53. Berkshire: CAB International.

LUCANSKY, T. W., AND K. T. CLOUGH. 1986. Comparative anatomy and morphology of
Asclepia~sperennis and Asclepia~s tuberose subspecies rolfsii. Botanical Gazette 147:
290-301.

LUKIANCHUK, J. L., AND S. M. SMVITH. 1997. Influence of structural complexity on the foraging
success of Trichogramnma minutum: a comparison of search on artificial and foliage
models. Entomolgia Experimentlis et Applicata 84: 221-228.

MALCOLM, S. B., AND L. P. BROWER. 1989. Evolutionary and ecological implications of
cardenolide sequestration in the monarch butterfly. Experientia 45: 284-295.

MAYHEW, P. J. 1997. Adaptive patterns of host-plant selection by phytophagous insects. Oikos
79: 417-248.

MYERS, J. H. 1985. Effects of physiological condition of the host plant on the ovipositional
choice of the cabbage white butterfly, Pieris rapae. Journal of Animal Ecology 54:
193-204.

NYLIN, S., AND K. GOTTHARD. 1998. Plasticity in the life history traits. Annual Review of
Entomology 43: 63-83.

ODENDAAL, F. J., Y. IWASA, AND P. R. EHRLICH. 1985. Duration of female availability and its
effect on butterfly mating systems. American Naturalist 125: 673-678.

,P. TURCHING, AND F. R. STERMITZ. 1989. Influence of host-plant densit and male
harassment on the distribution of female Euphydra~s anicia (Nymphalidae). Oecologia
78: 283-288.







































Figure 2-7. Controlled environment used to acclimate Monarch larvae to their host plant





I 1


Figure 2-1. Entrances to the three butterfly farms used throughout the experiment. Clockwise
from upper left, entrances to El Bosque Nuevo, Shady Oaks, and Greathouse butterfly
farms


Table 2-1. List of butterfly and host plants used
Butterfly Species for Costa Rica
Heliconius sapho
Caligo memnon
M~orpho peleides


Host plant Species for Costa Rica
Pa~ssiflora auriculata
Heliconia sp.
Leguminosa cristobal

Host plant Species for Florida
Asclepia~s cura~ssivica
Plantago lan2ceolata
Petroselinum criszur


Butterfly Species for Florida
Da~naus plexippus
Junonia coena
Pap~ilio pzolyxenes









throughout the entire range of transect stations 1 through 11. These data ultimately suggest that

butterflies are more abundant in the areas directly adjacent to the butterfly farm epicenter, and

are correspondingly less abundant farther from the farm epicenter. This idea will be discussed

more later, but for now, it is necessary to stress the fact that there was a marked difference

between transect stations 1 through 4, and transect stations 5 through 11.

The difference between 1 through 4 and 5 through 11 is likely due to the researcher' s

placement of the farm epicenter perimeter. This imaginary perimeter line was placed in order to

designate the perimeter of the actual working part of the farm. It was designed to incorporate all

the hypothesized attractive forces, such as host plants, nectar plants, rearing cages, butterfly

gardens, and laboratories that could potentially influence oviposition preference by containing

resources used by female butterflies. Because there were multiple areas with host and nectar

plants throughout each of the farm epicenters, some of the less attractive and less resource-

loaded components of the farm epicenters, such as closed laboratories and more shielded plant

nurseries, may have acted as a buffer between the perimeter and the more attractive rearing cages

and host and nectar plant nurseries. If wild populations of butterflies are being attracted

principally by these interiorly located plants and/or the associated butterflies, they may fly

directly to these areas and subsequently limit future flights to a set radius around these host and

nectar plants, while searching for alternatives depending on their oviposition preference.

Perhaps the 30 meter transect station (at station #4) approached the outer limit to their searching

behavior, rendering transects 5 through 11 less obvious. According to Renwick & Chew (1994),

learning plays a vital role in the recognition of suitable sites for oviposition. Thus, it could be

possible for butterflies to have learned the location of the best host plants within the farm

epicenter prior to the start of this experiment, having already assigned them oviposition priority.









farm had been in operation for the four-year period, but relative abundance rankings were not

significantly different among the species surveyed (Gordon and Ayiemba 2003). In summary,

butterfly populations increased proportionately during the first four years that the farm was in

place. Gordon and Ayiemba concluded that this was not harmful to the natural butterfly

populations since the original "balance" of common and rarer butterfly species in the local fauna

remained intact (Gordon and Ayiemba 2003).

Parasitism and predation at butterfly farms have also been investigated in previous

studies. Omenge (2002) reported that in small-scale butterfly farms in Kenya, most farmers

experienced noticeable predation and parasitism from ants, spiders, wasps, and lizards (Omenge

2002). Not surprisingly, Omenge (2002) found that farmers with more sophisticated butterfly

flight cages experienced fewer incidences of parasitism and predation than farmers with

makeshift, more rudimentary cages. He reasoned that this was due to the disintegration of the

screening on the poorly made and maintained flight cages, allowing parasites and predators to

have direct access to their host or prey (Omenge 2002).

Obj ectives

To further the study of butterflies in relation to butterfly farms, the present proj ect was

initiated to investigate three maj or aspects of butterfly life history, all with respect to the

butterflies' proximity to the central operational area of the butterfly farm, which will be referred

to hereafter as the farm epicenter. The first goal was to analyze butterfly abundance based on

oviposition. Contrary to previous studies, such as Gordon and Ayiemba (2003), which used

visual sightings of adults as indicators for abundances, my study used the deposition of eggs as

an indicator for butterfly abundance. While egg censusing will not measure the presence of

butterflies not laying eggs, such as males or non-gravid females, this method allows continuous

data collection over 24-hour periods, as new egg counts were recorded regularly every 24 hours.


















































-
r
i~ v i'~jc
;r. r .r
'' ~L~iL
.u ~


Figure 2-3. Dense jungle causing one row of host plants to be shifted slightly









BIOGRAPHICAL SKETCH

Court Whelan was born in July of 1983, in Clearwater, Florida. His love for nature and a

passion for living things began almost immediately and progressed through his participation in

nature programs, from summer camps to school field trips. His formal training in biology and

entomology began during the second semester of his undergraduate career at the University of

Florida (Spring 2002), upon taking a basic entomology course taught by Dr. Don Hall of the

department of Entomology and Nematology. After meeting with Dr. Hall to discuss his affinity

for entomology and the programs and advanced classes that were offered at UF, Court declared

his maj or as entomology immediately and has never looked back since. During his junior year of

undergraduate work (2004), he was offered a student internship with an ecotourism tour operator

to co-lead a group of 18 high school students on a 2-week field trip to the rainforests and coral

reefs of Belize. With the responsibility to instruct the students on tropical entomology and

ecology as well as co-lead the group throughout the trip, a strong passion for ecotourism

emerged simultaneously with a passion for tropical entomology and ecology. Following the trip,

his passion for ecotourism and entomology only grew stronger, as he immediately began

working with Dr. Thomas Emmel, Dr. Jaret Daniels and Dr. John Capinera on developing a new

graduate program entitled "ecotourism entomology." While continuing his undergraduate career,

he began working with a local tour operator, which specialized in tropical ecotourism, in

arranging and leading ecotourism trips. After receiving his Bachelor of Science degree in the

Spring of 2005, Court received an Alumni Fellowship from the University of Florida and was

admitted to graduate school in the Department of Entomology and Nematology. He received his

M. S. degree in 2008 and continues to arrange and lead ecotourism trips that are focused on

tropical biology, ecology and entomology while pursuing his Ph.D. in the newly formed

discipline of ecotourism entomology.








































Figure 4-2. More open layout of the area surrounding Greathouse farm. In the distance, there is
a residence with a butterfly garden a probable attractive factor









Factors Affecting Butterfly Dispersal

There are a number of different factors that influence a butterfly to either leave or remain

in a patch of habitat. Some factors are individual characteristics of the organism such as age or

sex (Kuussari et al. 1996) while others are characteristics of the habitat, such as the presence of

nectar sources and host plant density (Baker 1969, Shreeve 1995). The larval resource (host

plants) and adult resource (nectar plants, sap flows, rotting fruit) are important ecological factors

affecting movement in butterflies, as well as other holometabolous invertebrates (Brommer &

Fred 1999). Similarly, the sex of the butterfly is correlated with the rate of dispersal, as is the

insect's age. However, there is contention among experts as to the nature of the correlation

between sex and age with rates of dispersal (Kuussaari et al. 1996, Lawrence 1987). Although

these single factors may exhibit a certain degree of correlation with dispersal, one must question

whether correlation proves causation. As seen in Bowler and Benton (2005), many of these

single factors are associated with one unifying condition: conspecifie density.

Conspecifie density, which measures the density of organisms of the same species in a

given area, is seen to have a great influence on butterfly movement and dispersal (Dethier &

MacArthur 1964). While separate factors such as age, sex, and habitat resource availability may

play their individual parts, conspecifie density may be the most powerful, as it lumps many of

the possible factors that govern dispersal rates into one central theme. Food availability, sex

ratio, and parasite/parasitoid presence are all correlated to conspecific density (Bowler & Benton

2005). Therefore, understanding the effect that conspecifie density has on dispersal is useful in

understanding the entire system of dispersal. However, the correlation between conspecific

density and butterfly dispersal is not as definitive as, say, the correlation between nectar source

presence and butterfly movement into or out of a patch. With no surprise, there is a positive

correlation between butterfly movement into the patch and the presence of high numbers of









CHAPTER 2
MATERIALS AND IVETHODS

Oviposition as Index to Butterfly Abundance

This experiment occurred in a total of three geographic locations, with all measurements

and experiments set up identically in each location. The first two trials were conducted in north

central Florida, at Shady Oaks Butterfly Farm and Greathouse Butterfly Farm. The third trial

was conducted in northern Guanacaste Province, Costa Rica, at El Bosque Nuevo Butterfly Farm

(Figure 2-1).

For each location in this study, the researcher designated an "epicenter," as shown in

Figure 2-2. The epicenter contained all butterfly host plants being grown at the farm (either for

sale or for use in rearing and breeding), all butterfly cages, all rearing rooms, laboratories, and

all gardens. The farm epicenter is essentially the operational part of the farm. If there was farm

property not specifically used in the butterfly farming operation (such as access roads, parking

lots, etc.), it was excluded from the epicenter. The epicenter circumference was then measured

using a rolling-wheel electronic tape measure.

A total of nine rows of potted (one-gallon size) butterfly host plants were placed, starting

at the perimeter of the epicenter and radiating out to a distance of 100 meters for each row. Each

100-meter row was then divided further so that one host plant was placed at each 10-meter

interval. Therefore, 11 host plants were placed along each row (Figure 2-2). In order to

maintain the plants' health, a water saucer was placed under each pot to maximize water

retention. In order to distinguish the experimental plants from other nearby conspecifics, orange

marking flags were used to identify the experimental plants.

Each host plant row was evenly spaced and angled from the two adj acent rows as

accurately as possible so that no two rows were closer than other rows. Using the previous










BROMMER, J. E., AND M. S. FRED. 1999. Movement of the Apollo butterfly Pmarnssius apollo
related to host plant and nectar plant patches. Ecological Entomology 24: 125-131.

BROWER, L.. P 969. Ecological Chemistry. Scientific American. 220: 22-29.

1984. Chemical defense in butterflies. In R. I. Vane-Wright and P. R. Ackery (Eds.)
The Biology of Butterflies, pp. 109-134. Academic Press, London.

AND L. S. FINK. 1985. A natural toxic defense system: cardenolides in butterflies vs.
birds. In N. S. Braveman and P. Bronstein (Eds.) Experimental Assessments and
Clinical Applications of Conditioned Food Aversions, pp. 171-188. New York
Academy of Sciences, New York.

CLOBERT, J., E. CANCHIN, A. A. DHONDT, AND J. D. NICHOLS. 2001. Dispersal. Oxford
University Press, Oxford, UK.

COHEN, J. A., AND L. P. BROWER. 1982. Oviposition and larval success of wild monarch
butterflies (Lepidoptera: Danaidae) in relation to host plant size and cardenolide
concentration. Journal of the Kansas Entomological Society 55: 343-348.

COURTNEY, S. P. 1982. Coevolution of pierid butterflies and their cruciferous foodplants. IV.
Host apparency and Anthocharis cardamines~ddd~~~ddd~~~dd oviposition. Oecologia 52: 258-265.

AND J. FORSBERG. 1988. Host use by two pierid butterflies varies with host density.
Functional Ecology 2: 67-75.

G. K. CHEN, AND A. GARDNER. 1989. A general model for individual host selection.
Oikos 55:55-65.

AND T. T. KIBOTA. 1990. Mother doesn't know best: selection of hosts by ovipositing
insects. In E. A. Bernays (Ed.) Insect-Plant Interactions, pp. 161-188. CRC Press, Boca
Raton, Florida.

CRAIG, T. P., AND T. OHGUSHI. 2002. Preference and performance are correlated in the spittlebug
Aphrophora pectoralis on four species of willow. Ecological Entomology 27: 529-540.

DEMPSTER, J. P. 1983. The natural control of populations of butterflies and moths. Biological
Reviews 58: 461-481.

DETHIER, V. G. AND R. H. MACARTHUR. 1964. A field' s capacity to support a butterfly
population. Nature 201: 728-729.

DIXON, C. A., J. M. ERICKSON, D. N. KELLETT, M. ROTHSCHILD. 1978. Some adaptations
between Danaus plexippus and its food plant, with notes on Danaus chrysippus and
Euploea core (Insecta: Lepidoptera). Journal of Zoology 185 (4): 437-467.









CHAPTER 3
RESULTS

Main Experiment Butterfly Oviposition in North Central Florida

After a period of 10 days of data collection at two north central Florida butterfly farms,

60 observations (30 for each farm) were taken on oviposition for each of 11 transects. At these

two farms, sufficient data were collected from Monarch butterfly oviposition, but neither farm

yielded sufficient oviposition data for the Eastern Black Swallowtail or the Common Buckeye.

Although Buckeyes and Swallowtails were frequently seen at both farms by the researcher, no

Buckeye eggs and only 2 Swallowtail eggs were observed during the entire 10-day period on all

transects. Therefore, statistical models and summary statistics are meaningless for these two

butterflies. Monarch oviposition, however, was indeed commonplace, and therefore is the main

subj ect for this oviposition experiment in north central Florida.

Mean egg number per day was highest for those transects closest to the farm epicenter.

Then, egg number proceeded to decline and level out at a lower mean for transects more distal to

the farm epicenter (Figure 3-1). Figure 3-1 provides a summary of mean egg numbers per day at

each transect for monarch oviposition at the north central Florida farms. Summary statistics are

provided in Table 3-1. From these data, it is evident that the highest egg numbers are found

closer to the farm, with transects 1, 2, 3, and 4 yielding the highest egg numbers (3.033, 2.933,

2.933, and 2.700 eggs/day respectively). Starting with transect 5, which exhibited an average of

1.683 eggs/day, egg numbers then dropped off by approximately 1.5 eggs per day. Transects 5

through 7 exhibit a downward sloping trend as egg numbers proceeded to decline until rising

again at transect 8. However, transect 9 exhibited the lowest egg number of all, at 1.067

eggs/day. Then, a slight increase was exhibited with transect 10 seeing 1.4 eggs/day and transect

11 seeing 1.633 eggs/day. This up and down pattern between 5 and 11 is only slight when










(T-test at alpha = 0.005). An average of 58% larval survivorship on plants in the near location is

significantly greater than the 25% larval survivorship at the far location (Figure 3-8). Lastly,

larval survivorship between near and far locations in the control plot was found not to differ

significantly (Table 3-5).

Larval Experiment: Larval Survivorship as Determined by Distance from Farm Epicenters
in Costa Rica

Identical to the north central Florida farms experiment, a control plot was run in

conjunction with the experimental farm plot. After collecting the data, several comparisons were

made. First, using a T-test at alpha = 0.01 (larger alpha due to smaller sample size), it was

determined that there was no significant difference in total survivorship between control and

experimental plots (i.e., not accounting for distance from the epicenter). Percent survivorship for

the experimental plot was 83% versus 63% for the control plot (Table 3-3). The second

comparison was between near and far locations for larval survivorship in the experimental plot.

With near and far points exhibiting 90% and 75% larval survivorship, respectively, a T-test at

alpha = 0.01 concluded that there was no significant difference between the two distances (Table

3-4). However, a T-test at alpha = 0.01 did find a significant difference in larval survivorship

when comparing near and far treatments in the control plot (Figure 3-9). Percent survivorship in

this control plot was found to be 40% for the near plot and 85% for the far plot (Table 3-5).

Egg Parasitism

The egg parasitism study in north central Florida generated significant data. However,

the findings were very unexpected. By collecting a total of 12 eggs per transect, without regard

to day, none of the collected eggs yielded any evidence of parasitism. One hundred percent of

eggs collected from all transects hatched to produce live Monarch larvae. Due to the limited










many other studies of Lepidoptera have found that low survivorship of early instars is

responsible for most of each generation's mortality, a truly grave realization (Cohen & Brower

1982, Zalucki & Kitching 1982, Dempster 1983, Kyi et al. 1991). For these reasons, third instar

larvae were the subj ect for this part of the study. Third instars would not be as vulnerable to

catalepsies, while also having less of a proclivity to sequester high levels of cardiac glycosides.

Despite these studies, there are very few published studies that estimate natural predation

risks for butterfly larvae. However, one of these few studies found staggering results. Bernays

(1997) made a comparison of predation risks while Lepidoptera larvae were feeding versus when

they were resting. He found that predation levels while feeding increased three-fold for

Manduca sexta and 100-fold for Uresiphita reversalis. There is a great need for more research

to be done examining the predation risk for butterfly larvae in a natural environment.

Due to the extensive literature base, as well as their prevalence in butterfly farm cultures

and in the Florida wild, Monarch butterflies (Danaus plexippus) were selected as the principal

study organism for my experiment at north central Florida butterfly farms. Due to the

predominant presence and abundance of the Common Morpho (M~orpho peleides) at Costa Rica

butterfly farms, this Morpho was selected as the primary study organism at Costa Rica butterfly

farms.

Butterfly Parasitism

The relative importance of butterfly parasitism studies is being realized more and more as

butterfly captive breeding and ecotourism programs continue to emerge and grow. While the

maj ority of past studies have focused on the positive effects, or at least the theoretically positive

effects, of parasites and parasitoids of Lepidoptera that are used in biocontrol, the research and

findings on these situations are relevant and contribute to the overall understanding of

lepidopteran parasitoid behavior. Nevertheless, there is a need for additional research to be done









need for research that investigates the rate of egg parasitism within proximity to the butterfly

farm to critically examine this problem.

Dispersal

Generalities and supportive documentation derived from previous dispersal studies have

been difficult, mainly due to the amount of variation between the definitions of dispersal as well

as the great number of differences in spatial scales used in such studies. Perhaps the two most

common definitions for dispersal are (1) movement from the natal patch to the breeding patch,

and (2) movement between breeding patches (Clobert et al. 2001). However, these definitions

do not mention nor take into account the distances or even other reasons for dispersal, such as

foraging, for example. To avoid complicated semantic issues, this study will adopt a broad

definition of dispersal as stated by Bowler and Benton (2005), which defines dispersal as any

movement between habitat patches, and habitat patches as areas of suitable habitat separated in

space from other such areas, regardless of distance.

The consequences resulting from dispersal are great for both the individual actively

dispersing as well as the populations involved (i.e., the population that the individual is

immigrating to as well as the population that the individual is emigrating from). The act of

simply moving from one patch to another may affect individual fitness, population dynamics,

and genetics, as well as species distributions (Dunning et al. 1995, Hanski & Gilpin 1997,

Hanski 1999, Clobert et al. 2001). The well-founded link between dispersal and population

dynamics necessitates a better understanding of the patterns, causes and consequences of

dispersal. This is critical to our overall understanding of population management and our ability

to predict population responses to changes in the environment (Bowler & Benton 2005).









the farm epicenter, much like the first experimental arrangement described earlier. However,

since there were only two host plants per row, one was placed at the edge of the epicenter

perimeter (distance = 0 meters) and the other was placed 100 meters away (distance = 100

meters) (Figure 2-6).

For each plant at the north central Florida location, five 3rd-instar Monarch butterfly

larvae were directly placed on the plant and allowed to feed in a controlled laboratory

environment for 24 hours (Figure 2-7). This step was to acclimate the larvae on the host plant

and to make sure that they were not prone to wander off of the plant within the 24 hours that the

trial was actually running. After this 24 hour acclimation period, the plants were taken from the

laboratory and placed in their appropriate positions in the field. After another 24 hours, data

were collected on the number of larvae remaining for each plant. This entire process was

replicated twice for each north central Florida location.

For the Costa Rica location, two rows of host plants of the Common Morpho (M~orpho

peleides) were placed in an identical fashion as the plants at the north central Florida location

(Figure 2-6). Because the host plants were trees and relatively large, they would not be able to

fit into a controlled laboratory room, as the Monarch larvae were. Therefore, an alternate

method was used. With the potted host plants already in the field, five 3rd instar Common

Morpho larvae were placed directly on the plant and then enclosed in a "predator screen" (Figure

2-8). This screen was a fine mesh bag tied at both ends (using a secure overhand reef knot), both

preventing the larvae from wandering as well as excluding parasites and predators from attacking

the larvae while they were being acclimated to the host plant (Figure 2-9). These screens are

used in the very same way by the butterfly farm and have provided successful results throughout









In this set of experiments examining butterfly abundance and ovipostion around north

central Florida butterfly farms, there were three main effects: day, location, and transect. The

following sections will discuss possible explanations for these results in detail.

Main Effect 1: Day

The effect that the number of days host plant exposure to gravid females in the vicinity

had on egg number throughout the experiment was very significant (p-value < .0001). With days

four, five, and six exhibiting the most oviposition, it is plain to see that butterfly abundances

were highest during the middle of the experiment, forming a bell curve. The most probable

explanation for this initial spike is that the first few days were required for the butterflies to

locate, accept, and learn the location of the host plants, which resulted in high oviposition levels

on the subsequent days. Then, once the host plants became "heavily" used and environmental

conditions stressed the plants, they became less suitable to gravid females, resulting in less

frequent oviposition. Butterflies rely on specific sensory cues in searching for appropriate host

plants, with visual perception playing a major role (Prokopy & Owens 1983, Singer 1993).

Using visual perception, butterflies overwhelmingly use the shape and color of the host plant as a

search tool, with the color and shape of leaves particularly important (Stanton 1984). According

to Courtney and Forsberg (1988), landing frequencies by gravid female butterflies have been

seen to depend on the relative abundance of the host plant in the area. With a wealth of

competing host plants in the general vicinity, all providing similar visual cues, it is possible that

some gravid females did not detect the experimental host plants immediately. As these

experimental host plants were detected by more and more gravid females, the oviposition rate

increased until reaching a peak -- in the middle of the experiment.

Once the plants elicited the maximum oviposition rate during the temporal median of the

experiment, oviposition began to decline sharply. It is possible that during the days that these










(provided that there is sufficient food supply available). Theoretically, a faster growth rate

results in less time spent in the vulnerable juvenile period, reducing the amount of time the

larvae is exposed to threats, and increasing the probability that the immature organism will

survive until reaching sexual maturity (Stearns 1992). However, Arendt (1997) suggests that

oftentimes organisms grow at a slower rate than they are physiologically capable. If it is indeed

true that a shorter juvenile period would lead to greater overall success of the animal (i.e., less

time susceptible to predation, parasitism, and other threats), why wouldn't the animal attempt to

maximize its growth rate? The answer is relatively simple. Growth rate and predation level are

both functions of activity level in many organisms (Werner & Anholt 1993). Commonly,

organisms that exhibit a maximum growth rate also require maximum nutrition and, therefore,

devote more time to foraging activity, which puts the animal out in the open more often and

increases its susceptibility to predation. Consequently, larval growth patterns are likely to be the

result of making "strategic decisions" where the organism weighs the costs and benefits of

various growth traj ectories and chooses an optimal rate dependent upon its particular

environment (Abrams et al. 1996, Nylin & Gotthard 1998, Gotthard 1999). In a laboratory

experiment, Gothard (2000) found that there is indeed a trade-off between butterfly larvae

growth rate and predation risk, using the butterfly Pararge~PPP~~~~PPP~~~PPP aegeria (Nymphalidae: Satyrinae).

When comparing a "fast-growing larvae" and "slow-growing larvae", they found that the daily

predation risk was 30 % higher in their "faster-growing larvae" treatment (Gothard 2000).

Therefore, we see that larval predation may be more of a problem for species that exhibit

naturally high growth rates.

Butterfly larvae are at particular risk to predation due to their feeding habits. They are

essentially feeding machines that devote their entire juvenile lives, as well as body plans, to









be too far developed, or perhaps if it has already been parasitized, the wasp will abandon the egg

and move on (Klomp & Teerink 1962).

These wasp parasites can be costly problems for Lepidoptera, a fact which has deemed

them powerful bio-control agents on pest Lepidoptera in agriculture. Unfortunately, they also

parasitize non-pest Lepidoptera that can thereby cause costly problems for butterfly farmers. As

mentioned earlier, exclusionary measures are frequently used to protect butterfly stock from

unacceptably high levels of parasitism. However, these measures are typically employed only in

special labs and buildings for butterfly eggs and other immature stages. Trichogramnma wasps

still present problems in other areas of the butterfly farming operation, such as outdoor plant

nurseries and gardens.

Parasitoids and the Farms

A common sight at butterfly farms is host and nectar plants growing in relatively vast

numbers around the property. Some of these plants are found under the cover of screens in

shadehouses, while others are out in the open to take advantage of sunlight as well as minimizing

overhead costs to the farm. In both cases, these plants have the potential to attract parasitoids,

creating a risk for the farm. As plants in the wild serve as resources to butterflies, so do plants

that are found in the open areas at butterfly farms and breeding centers. The potential for high

densities of butterfly eggs at these host plant nurseries would naturally attract parasitoids,

especially Trichogramnma wasps. Additionally, the close spacing of plants, common in these

butterfly farm nurseries, makes it very easy for parasitoids to walk and jump to adj acent plants

(Romeis et al. 2005). Research suggests that this augments parasitoid dispersal, which could

make these areas prime habitats for Trichogramnma wasps and other egg parasitoids (Surverkropp

1997). If this is indeed true, it means that butterfly farms could be supporting high populations

of parasitoids in the same general area as where they harvest and store butterfly eggs. There is a










LIST OF REFERENCES


ABRAMS, P. A., O. LEIMAR, S. NYLIN, AND C. WIKLUND. 1996. The effect of flexible growth rates
on optimal sizes and development times in a seasonal environment. American
Naturalist 147: 381-395.

ANDOW, D. A., AND D. R. PROKRYM, D. R. 1990. Plant structural complexity and host-finding by
a parasitoid. Oecologia 82: 162-165.

ARENDT, J. D. 1997. Adaptive intrinsic growth rates: an integration across taxa. The Quarterly
Review of Biology 72 (2): 149-177.

ATKINSON, D. 1994. Temperature and organism size a biological law for ectotherms?
Advances in Ecological Research 25: 1-58.

ATKINSON D., S. A. MORLEY, D. WEETMAN, AND R. N. HUGHES. 2001. Offspring size responses
to maternal temperature in ectotherms. In D. Atkinson and M. Thorndyke (Eds.).
Environment and Animal Development: Genes, Life Histories, Plasticity, pp. 269-286.
BIOS Scientific Publishers, Oxford, UK.

AVELAR, T. 1993. Egg Size in Drosophila: Standard Unit of Investment or Variable Response to
Environment? The Effect of Temperature. Journal of Insect Physiology 39 (4): 283-
289.

AZEVEDO, R. B. R., V. FRENCH, AND L. PARTRIDGE. 1996. Thermal evolution of egg size in
Drosophila melan2oga~ster. Evolution 50: 2338-2345.

BAGUETTE, M., I. CONVIE, AND G. NEVE. 1996. Male density affects female spatial behavior in
the butterfly Proclossiana eunomia. Acta Oecologia 17: 225-232.

BAGUETTE, M., C. VANSTEENWEGEN, AND I. CONVI. 1998. Sex-biased density-dependent
migration in a metapopulation of the butterfly Proclossiana eunomia. Acta Oecologia
19: 17-24.

BAKER, R. R. 1969. The evolution of the migratory habit in butterflies. Journal of Animal
Ecology 38: 703-746.

BERNAYS, E. A. 1997. Feeding by Lepidopteran larvae is dangerous. Ecological Entomology
22: 121-123.

BOPPRE, M. 1983. Leaf-scratching-a specialized behavior of danaine butterflies (Lepidoptera)
for gathering secondary plant substances. Oecologia 59: 414-416.

BOWLER, D. E., AND T. G. BENTON. 2005. Causes and consequences of animal dispersal
strategies: relating individual behavior to spatial dynamics. Biological Review 80: 205-
225.











5.0
4.5
4.0 -
3.5-
3.0
2.5
2.0
1.5-
1.0 -
0.5
0.0
1 2 3 4 5 6 7 8 9 10 11

SEgg Number 14.600 14.100 4.233 4.3001 2.6331 1.5331 1.867 2.4671 1.933 2.500 3.067
Transect (at 10m intervals)


Figure 3-2. Mean egg number at each transect for Greathouse Butterfly Farm


Transect (at 10m intervals)


Figure 3-3. Mean egg number at each transect for Shady Oaks Butterfly Farm










permits, etc. Additionally, these plants are often kept outside in order to maximize sunlight and

minimize overhead cost. Combining these three variables, there is a high chance that native

butterfly populations in the area will find the farm suitable as an area in which to oviposit on the

host plants, or to feed on the nectar sources.

Although this situation seems like it might be no more harmful than a butterfly garden,

there are several key differences. Firstly, these plants are frequently moved, pruned, and even

"de-egged" (eggs are removed just before using them in the butterfly enclosures to safeguard

against the introduction of egg parasites in quarantined zones). These disturbances could cause

the eggs to fall off or be removed from the plant. The second way that these nurseries differ

from a butterfly garden is that the host plants are often very dense and clustered together in a

homogenous pattern. This enables parasites to easily walk and jump to different plants (Romeis

et al. 2005). Butterfly gardens are typically more heterogenous in the way the plants are

arranged, serving as natural defenses to parasitoids. Lastly, predators such as vespid wasps, ants,

lizards and birds are likely to exploit a predictable source of food in these nurseries, predating

upon adults and larvae. It is clear that if the conditions are right, native butterfly populations

could be in danger of falling into an ecological trap.

From an environmental standpoint, the presence of an ecological trap could be very

detrimental, and the problem requires further investigation. However, it is not just an

environmental problem; the farm could also be negatively impacted as well. In the case of both

parasitoids and predators, high levels could pose problems for the farm's butterfly stock.

Predators and parasitoids may attempt to enter breeding and rearing centers, causing significant

economic loss to the farm.










DUNNING, J. B. J., D. J. STEWART, B. J. DANIELSON, B. R. NooN, T. L. ROOT, R. H. LAMBERSON,
AND E. E. STEVENS. 1995. Spatially explicit population models: current forms and
future uses. Ecological Applications 5: 3-11.

DUSSOURD, D. E. 1990. The vein drain; or how insects outsmart plants. Natural History 90: 44-
49.

EHRLICH, P. R., R. R. WHITE, M. C. SINGER, S. W. McKECHNIE, AND L. E. GILBERT. 1975.
Checkerspot butterflies: A historical perspective. Science 188: 221-228.

ERNSTING, G., AND J. A. ISAAKS. 1997. Effects of temperature and season on egg size, hatchling
size and adult size in Notiophihts biguttatus. Ecological Entomology 22 (1): 32-40.

FATOUROS, N. E., G. B. KIss, L. A. KALKERS, R. S. GAMBORENA, M. DICKE, AND M. HILKER.
2005. Oviposition-induces plant cues: do they arrest Trichogramnma wasps during host
location? Entomologia Experimentalis et Applicata 115: 207-215.

FEENY, P. 1991. Chemical constraints on the evolution of swallowtail butterflies. In J. Wiley
(Ed.) Plant-Animal Interactions: Evolutionarey Ecology in Tropical and Temperate
Regions, pp. 315-340. New York.

FISCHER K., P. M. BRAKEFIELD AND B. J. ZWAAN. 2003. Plasticity in butterfly egg size: why
larger offspring at lower temperatures? Ecology 84: 3138-3147.

A. N. M. BOT, P. M. BRAKEFIELD, AND B. J. ZWAAN. 2006a. Do mothers producing
large offspring have to sacrifice fecundity? Journal of Evolutionary Biology 19: 380-
391.

S. S. BAUERFEIND, AND K. FIEDLER K. 2006b. Temperature-mediated plasticity in egg
and body size in egg size-selected lines of a butterfly. Journal of Thermal Biology 31:
347-354.

FORISTER, M. L. 2004. Oviposition preference and larval performance within a diverging lineage
of lycaenid butterflies. Ecological Entomology 29: 264-272.

Fox, C. W., AND M. E. CZESAK. 2000. Evolutionary ecology of progeny size in arthropods.
Annual Review of Entomology 45: 341-369.

Fox, R. M. 1966. Forelegs of butterflies I. Introduction: chemoreception. Journal of Research on
the Lepidoptera 5: 1-12.

GATES, J. E. AND L. W. GYSEL. 1978. Avian nest dispersion and fledging success in field-forest
ecotones. Ecology 59: 871-883.

GILBERT, I. E. AND M. C. SINGER. 1973. Dispersal and gene flow in a butterfly species.
American Naturalist 107: 58-72.










and job opportunities to the community involved. By measuring butterfly abundance rates,

scientists are able to better understand the ecological dynamics and influences of a butterfly farm

and its impact on the surrounding areas. Should irregular patterns of butterfly distribution and

abundance occur (i.e., butterflies found to be not evenly dispersed throughout the farm property),

it was the obj ective of this study to not only observe and record such patterns, but also to provide

interpretation of the factors influencing such irregularity. More importantly, some investigators

have suggested that the farm may act as an ecological trap, where the butterflies are being

attracted to the farm epicenter under the auspices of reward for homing in on an environmental

cue (Schlaepfer et al. 2002). Ecological traps will be discussed further in a later section starting

on page 29. Finally, at a conservation level, understanding the reasons for patterns of parasitism,

predation and butterfly abundance in areas surrounding the farm will allow for better

management of the farm. Ultimately the hope would be that the results of this study might

provide supportive evidence that will minimize any potential negative influence the farm may

have on its environment, and result in a more stabilized, productive, and healthy ecosystem.

This, in turn, will produce a healthier farm ecosystem and likely increase the sustainability and

productivity of the butterfly farm and the surrounding human communities that may be

economically dependent on the farm.

The remainder of this introduction will detail the maj or topics addressed in this study,

including larval predation, egg parasitism, oviposition, dispersal, and ecological traps.

Larval Predation

The fitness of most organisms is closely related to the organisms' age and size at sexual

maturity (Steamns 1992). Reaching an optimal combination of age and size as an adult is largely

governed by how an organism performs in its juvenile stages, particularly concerning growth rate

(Steamns 1992). Regarding juvenile survivorship, one can see the benefits of a rapid growth rate










An initial hypothesis of the researcher was that butterfly farms may be acting as

ecological traps, presenting adult butterflies with seemingly suitable resources and habitat, only

to have their eggs parasitized and larvae predated upon at higher than normal rates. Although it

was evident that butterflies were in higher numbers surrounding the farm, larval predation was

less of a threat at the farm and especially low quite near the farm epicenter. Therefore, this study

suggests that butterfly farms may, in fact, not serve as ecological traps and in a way could have

the opposite effect by acting as a source for metapopulations in fragmented landscapes.

Although this is a relatively unsubstantiated assumption, it could provide the framework for an

intriguing follow-up study.

Being a preliminary study of butterfly abundance, larval predation, and egg parasitism at

butterfly farms and ecotourism areas, this set of experiments yielded interesting and helpful

results. Many possible explanations for trends in these data involve butterfly activity from

within the epicenter as well as possible attraction to the epicenter. Further studies should

concentrate on the specific areas in question, such as nurseries, outdoor gardens, and breeding

and rearing centers in order to conclusively evaluate specific attractive qualities they may

possess.







































Figure 4-5. Student groups participating in ecotourism may be perceived as threatening by larger
butterfly predators, such as birds and lizards









butterfly farm epicenter (Figure 4-3). At this residence was a butterfly garden. Filled with

abundant nectar sources and covering an area of approximately 200 square meters, this garden

was a butterfly attractant, as observed by the researcher. Although only one row of host plants

traversed this area, high rates of oviposition was observed for those transects proximal to the

garden in this particular row. Although this residence and garden were very close to the farm

epicenter, they could not be counted as part of the epicenter because it was not involved in the

operations at the butterfly farm.

One more thing to note about the Greathouse trial was the presence of a large monarch

butterfly enclosure, filled with hundreds of milkweed plants and adult and larval Monarch

butterflies located at the very edge of the farm epicenter (Figure 4-4). It just so happened that

one of the Monarch butterfly host plant rows used in this experiment was placed directly

adj acent to this enclosure, as dictated by the randomization scheme designed prior to the start of

the experiment. Not coincidentally, the first three host plants radiating out from the epicenter

here also exhibited the highest number of Monarch eggs throughout the entire experiment, with

some days seeing upwards of 20 eggs on each plant. High oviposition levels here are possibly

the result of high plant visibility, and may have been attracting gravid females. When the

females arrive to oviposit on what they feel will be suitable habitat, they are excluded from the

enclosure and choose to oviposit elsewhere. With several of the experimental host plants located

in close range, they become a probable choice for the gravid females. Although these plants did

have relatively high levels of oviposition, these transect stations were not outliers in terms of

overly high levels of oviposition when compared to the other data from this study. Oviposition

rates were affected by the number of days the plants at these transects were exposed to gravid



































Figure 2-5. Screened vials and stereoscope used to observe the emergence of parasites or larvae


Figure 2-6. Diagram of larval predation experiment











50
40
30
20
10
0

% Survivorship


% Survivorship 58 25
Distance


Figure 3-8. Mean % survivorship among larvae at far and near distances in the experimental plot
for north central Florida farms


Control Treatment Experimental Treatment
14 41
Treatment


Figure 3-7. Mean % survivor ship among larvae at experimental and control plots for north
central Florida farms


I


Near










literally hundreds ofAnartia over the course of five days, not a single egg was found on the host

plants.

After the experiment was complete, the researcher began to compare the north central

Florida test versus the Costa Rica test in order to reason why there was such a drastic difference

in oviposition levels. There was indeed one glaring difference between the trials. In both north

central Florida and Costa Rica, orange marking flags were used to delineate transects and mark

the experimental host plants. It could have been possible that Monarch females saw the orange

marking flag and were attracted to it as being a possible mate, subsequently ovipositing on the

plant, or returning at a later time to oviposit. At this point, shiny blue plastic material was used

to mimic butterfly presence in the same way the orange flags were presumed to do. After the

blue plastic was tied to the host plants, the 10-day experiment was repeated (Figure 4-7). Again,

there was no oviposition. Although the orange marking flags could have served as an attractant

in the north central Florida trials, there is no proof.

This brings us to the question of why Monarch butterflies were the only species that

oviposited during the north central Florida trials? Although Swallowtails and Buckeyes were

seen regularly in and around the farms, there was virtually no oviposition (there were two eggs

seen of the Swallowtail during the entire experiment). There are two main factors believed to be

the cause for this lack of oviposition. The first is a relative lack of visual conspicuousness.

Courtney (1982) suggested that the level of visual conspicuousness (i.e., how obvious a plant is

compared to its surroundings) directly affects oviposition in butterflies. Specifically, as visual

conspicuousness increases, so does oviposition. Based on the size, color, and leaf shape of the

two host plants used for the Swallowtail and Buckeye (parsley and plantain, respectively), they

likely exhibited a relatively low level of visual conspicuousness in relation to their surroundings.










larger, while the total number of eggs she lays is lessened. These initial beliefs have been

confirmed in many recent studies (Azevedo et al. 1996; Yampolski & Scheiner 1996; Atkinson

et al. 2001; Fischer et al. 2003, 2006a, 2006b; Steingenga & Fischer 2007) and could provide

insight into butterfly emergence and dispersal patterns. This plasticity between egg size and

clutch size is not an even trade-off, however, as total reproductive investment must be increased

at higher temperatures (Avelar 1993; Ernsting & Isaaks 1997, 2000; Fischer et al. 2003). For

example, lifetime fecundity was almost twice as high at 270 C than at 200 C in a study by Fischer

et al (2003). Although the mother' s biological response of egg size and clutch size to

temperature was gradual (two to three days of cold could alter egg laying behavior) in Fischer et

al.'s study (2003), it was also found to be reversible after two to three days of warm weather.

We can surmise from these studies that oviposition will be in greater numbers in warmer

environments than in colder environments.

Oviposition Preference

Most butterflies are oligophagous. This habit of feeding on host plants from a very

limited number of plant families, oftentimes only a single family, suggests that there may be

basic similarities within plant families that signal or cue a butterfly to oviposit (Feeny 1991).

The host plant may be attracting the butterfly by exuding a chemical, visual, or tactile cue. In

any case, it is the j ob of the adult butterfly to choose the best host plant for its soon-to-be

offspring. Larvae of many phytophagous and holometabolous insects are to a certain degree

immobile and are not able to search for appropriate food over comparatively large distances

(Rausher 1979a). Females that oviposit on plants with low larval survivorship or poor larval

growth will likely leave fewer descendents than females that oviposit on the more suitable host









It was the goal of this study to conduct a survey of butterfly abundances in relation to

their proximity to the butterfly farm. Should wild butterflies be significantly more abundant in

areas closest to the farm, this may suggest that they are indeed being attracted to the farm.

However, additional studies are necessary to either dismiss this as insignificant or to verify this

attraction and distinguish this factor from the possibility that escaped larvae or adults are

artificially inflating population numbers on the outskirts of the farm. In addition, larval

predation and egg parasitism also need to be examined in this study as a function of proximity to

the butterfly farm in an effort to document all facets of a possible ecological trap.

Oviposition

The evolutionary process of insect egg laying, or oviposition, is a topic that has received

a significant amount of attention in past studies. As with other insects, oviposition in butterflies

has been studied from many different angles. Those pertaining to variations in egg and clutch

size, oviposition preference and oviposition habitat have particular application to this study.

Variation in Egg and Clutch Size

Egg size is a valuable measurement when studying insects, as it is closely related to

fitness (Fox & Czesak 2000, Fischer et al. 2003). In an ultimate sense, it involves how mothers

choose to allocate resources among their offspring. Almost always, as energy available for

reproduction is held relatively constant, a larger egg size is correlated with a smaller clutch size,

and vise versa (Smith & Fretwell 1974). However, it is not clearly known how a clutch is

optimized by selection for either numbers, or for individual egg size. Although it may not be a

panacean explanation for why any gravid female insect would choose greater egg size over

greater clutch size, one study proposed that there is a correlation between temperature and egg

size (Atkinson 1994). Atkinson (1994) coined the term temperature-size rule, which essentially

states that as the temperature experienced by the mother is lowered, the size of her eggs become









nectar sources (Shreeve 1995). Nevertheless, understanding the effects that conspecific density

has on dispersal is key in this study of butterfly abundances at butterfly farms, which have

artificially high conspecific densities due to their butterfly stock and/or dense planting of nectar

sources inside and outside the cages.

Conspecific density

Previous studies have observed that individuals disperse at a higher rate away from an area

when the area exhibits a low conspecific density, suggesting a negatively density-dependent

dispersal (Gilbert & Singer 1973; Kuussaari et al. 1996, 1998), while other observations find that

individuals disperse at higher rates away from areas exhibiting high conspecific density,

suggesting a positively density-dependent dispersal (Dethier & MacArthur 1964; Odendaal et al.

1989; Baguette et al. 1996, 1998). The next two sections will concentrate on the differences

between these two concepts and common arguments in favor of one or the other.

Negatively density-dependent dispersal

Negatively density-dependent dispersal is a term used to describe a situation in which

individuals disperse at a higher rate away from an area when the area exhibits a low conspecific

density (Gilbert & Singer 1973; Kuussaari et al. 1996, 1998). According to Hanski et al. (1994)

and Kuussaari et al. (1998), this could be the result of mate scarcity, as a butterfly inhabiting an

area with low conspecific density would have a lesser chance of finding an appropriate mate and

may choose to leave the patch in order to hopefully increase chance of finding a mate elsewhere.

The concept of high disperal rates away from areas of low conspecific density gave rise to the

conspecific attraction hypothesis, which states that high density populations will result in higher

immigration rates to and lower emigration rates away from the habitat in question (Ray et al.,

1991). There are two supporting arguments to this hypothesis. The first, given by Odendaal et

al. (1989), is that butterflies are attracted to and ultimately remain in habitats with high


























Distance

Figure 3-9. Mean % survivorship among larvae at far and near distances in the control plot for
the northern Costa Rica farm










4-2 More open layout of the area surrounding Greathouse farm ..........._... ...._._..........77

4-3 A residence in the area surrounding Greathouse farm, complete with butterfly garden ...78

4-4 Large screen enclosure housing hundreds of milkweed plants ................. ............... ....79

4-5 Student groups participating in ecotourism may be perceived as threatening by larger
butterfly predators, such as birds and lizards ................. ...............80........... ..

4-6 Heliconia plants around the exterior of the farm epicenter could have acted as
competition for oviposition in the area ....._ .....___ ........__ ...........8

4-7 Blue plastic material used in hopes of stimulating oviposition by M~orpho peleides ........82







































Figure 2-2. Diagram of farm epicenter (green circle) with radiation rows of hostplants at 10
meter transects (yellow dots). Transects repeated for each row, but shown here for
only one row. Three different host plants were used, corresponding with the three
differently colored rows in this diagram










compared to the drop from 1 to 5. From these data, there appears to be a noticeable difference in

egg numbers between transects in regards to distance from the farm epicenter.

Looking at the data summarized in Table 3-1, we see that the highest standard deviations

correspond with the highest mean egg numbers. These high standard deviations wreak havoc on

statistical tests, lessening any significance that the data may inherently possess. Consequently,

we cannot say that there is a significant difference in egg averages between transect 1 and

transect 11, much less between other more closely-spaced transects. However, when we look at

the 3rd quartile data (Table 3-1), there is a clear increase in frequency of higher egg numbers with

transects more proximal to the epicenter, specifically transects 1 through 4. Additionally,

median values are significantly higher with transects 1 through 3. Perhaps due to the sample

size, we cannot prove significant differences between any of the transects using summary

statistics alone. The next section will detail more complicated statistical analyses and findings.

In order to further analyze butterfly oviposition, the researcher used the model of

repeated measures over time, treating the experiment as a randomized block design with day,

location and transect as the factors. This allows the researcher to put the fixed effects of

location, day and transect into blocks in order to measure how influential each fixed effect is to

the variance in egg number. The raw data were then transformed (i.e., egg numbers) using a

log+1 transformation in SAS to standardize the residuals in order to meet the assumptions of the

model. The transformed data were then analyzed in SAS using a proc mixed procedure, which

generated the significance statistics found in Table 3-2.

Variations in egg number can be attributed to 4 different effects: location, transect, day,

and the interaction between location and day. There were no reported interaction effects

between location and transect, nor day and transect. Table 3-2 shows a list of the effects and









measurement of circumference, the researcher was able to calculate equal distances and equal

angles for the spacing between all rows.

A total of three species of host plants were used at each butterfly farm, with each

intended to elicit oviposition from a specific species of butterfly. See Table 2-1 for the complete

list of host plants and butterflies. Different species of host plants were used for a different set of

butterflies at the Costa Rica butterfly farm (Table 2-1), but for both north central Florida

locations the same plants were used. To recapitulate, one plant was placed at each 10m transect

for each row and there were nine rows of host plants total. With three different host plants per

farm, each host plant was replicated three times so that there were three rows for each plant.

Each row contained 11 plants, from a distance of zero (adj acent to the epicenter perimeter) to a

distance of 100 meters. Therefore, 99 individual plants were used for each farm.

Challenges in placing the plant rows presented themselves in the experiment. First, each

farm was situated within 100 meters from a road. Therefore, several plant rows had to cross

these roads at which points the researcher preserved the 10-meter spacing as much as possible.

In the case of Costa Rica, dense jungle made 10-meter spacing difficult with two of the nine

rows, and as a result the angle between them was slightly more acute than the other rows in this

situation (Figure 2-3).

In the case of the north central Florida farms, the plants were placed on the edges of roads

(Figure 2-4), which caused several transect intervals to be slightly longer than 10 meters.

Another problem with having roads so close to the farm being studied was that several plants

placed near the road were stolen. However, this happened only once, and the plants were able to

be replaced on the following day, thus minimizing the loss of data.












TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES ............ ...... .__ ...............7...


LIST OF FIGURES .............. ...............8.....


AB S TRAC T ............._. .......... ..............._ 10...


CHAPTER


1 INTRODUCTION ................. ...............12.......... ......


Butterfly Farming and Ecotourism ................ ...............12................
Obj ectives ................. ...............14.......... .....
Larval Predation............... ...............1
Butterfly Parasitism .............. ...............19....
Parasitoid Behavior .............. ...............20....
Parasitoids and the Farms ................. ...............22................

D ispersal ............... ................ .... .........2
Factors Affecting Butterfly Dispersal .............. ...............24....
Conspecific density ................ ........ ..............2
Negatively density-dependent dispersal ................. ...............25........... ....
Positively density-dependent dispersal .............. ...............26....
Dispersal and the Farms .............. ...............27....
Ecological Traps ................ ...............28.................
Oviposition .............. .. .. .......... ... .........3
Variation in Egg and Clutch Size .............. ...............30....
Oviposition Preference ........._.... ...............3.. 1..............

2 MATERIALS AND METHODS .............. ...............34....


Oviposition as Index to Butterfly Abundance .............. ...............34....
Egg Parasitism .............. ...............36....
Larval Predation............... ...............3


3 RE SULT S .............. ...............47....


Main Experiment Butterfly Oviposition in North Central Florida ........._.... ........_......47
Main Experiment: Butterfly Oviposition in Costa Rica .............. .. ... ...............50.
Larval Experiment: Larval Survivorship as Determined by Distance from Farm
Epicenters in North Central Florida ............... .. .. .... ...... ... ......... ... .........5
Larval Experiment: Larval Survivorship as Determined by Distance from Farm
Epicenters in Costa Rica ................ ...............51................
Egg Parasitism .............. ...............51....










LIST OF TABLES


Table page

2-1 List of butterfly and host plants used ................. ...............39........... .

3-1 Summary statistics for North central Florida oviposition experiment ............... .... ...........53

3-2 Significance values obtained from the SAS proc mixed procedure (c=0. 1) ........._._.........53

3-3 Mean % survivor ship among larvae at experimental and control plots for all
geographic locations .............. ...............56....

3-4 Mean % survivor ship among larvae at far and near distances in the experimental plot
for all geographic locations............... ...............5

3-5 Mean % survivor ship among larvae at far and near distances in the control plot for
all geographic locations .............. ...............56....









The host plants used in this experiment were the same species used by the farm for

breeding their own butterfly stock. However, these host plants could have been the wrong

choice, nonetheless. Landing frequencies by gravid female butterflies, and ultimately

oviposition rates, have been seen to fluctuate with the seasonal changes in host plant abundances

(Rausher 1979a). It is possible that while farm-raised butterflies will oviposit on these host

plants year-round, wild butterflies may be seasonally adapted to searching for different species

during the time this experiment was conducted. This could also apply to the north central

Florida experiment, providing explanation for why there was little to no oviposition by both

Eastern Black Swallowtails and Common Buckeyes.

The layout of the Costa Rica farm posed a particular problem for measuring oviposition

of Caligo memnon. Near the farm epicenter, but outside of the operational part of the farm, there

was a huge stand ofHeliconia plants. These were the same species used at the farm to rear their

own Caligos and consequently, the same species used in this experiment (Figure 4-6). With

hundreds of Heliconia plants very near two of the experimental Heliconia rows, the hostplant

competition here for visiting female Caligo was especially high.

Throughout the beginning phases of the experiment, there was a very high abundance of

Anartia fatima butterflies around the farm. As the experiment progressed, and day after day no

eggs were found at any of the host plant transects stations, several host plants for Anartia fatima

were placed by the researcher around certain nectar sources and other areas that they frequented.

Thinking that perhaps the wrong indicator species was chosen to test for abundance (although

M~orpho peleides sightings were common), the researcher began to monitor these Anartia host

plants in addition to the other experimental plants. Despite being within an arm's reach of








































Figure 2-4. Several plants needed to be placed close to roads in the Florida experiments









CHAPTER 5
DISCUSSION

Throughout the world today, butterfly farming is becoming increasingly common as a

source of revenue or commerce for communities in both developed and developing nations. It

has been reported that this industry can generate six-digit annual revenues in even the most rural

of communities in developing nations (Gordon & Ayiemba 2003). While specializing in

butterfly rearing and breeding, many of such farms also provide educational opportunities for the

surrounding communities, which serve to educate the public on farming practices as well as

butterfly and habitat conservation. By providing economic incentives to the local community

and actively engaging in nature education and conservation, butterfly farming is rapidly

becoming a part of ecotourism and conservation efforts across the globe.

With the emergence of butterfly farming operations, so too emerges a need to better

understand factors affecting butterfly natural history around the farm. It was the objective of this

study to better understand the impacts of farms on butterfly abundances, using oviposition as an

indicator, butterfly larvae predation rates, and butterfly egg parasitism, all with regard to

proximity to butterfly farms.

Butterfly Abundance and Oviposition at North Central Florida Farms

Monarch butterfly oviposition was, on average, in greater numbers at transect stations 1

through 4 when compared to transect stations 5 through 11. Although there is not a smooth

downward sloping trend in egg number as one goes from transect stations 1 to 11, we can say

that Monarch butterflies are much more likely to be found ovipositing on plants within 40 meters

of the butterfly farm epicenter. When examining transect stations 1 through 4, specifically, we

observed a completely negative correlation between egg number and transect station number as

one goes from transect station 1 to transect station 4, which was hypothesized to occur









contributed critical knowledge to our understanding of these organisms. Klomp and Teerink

(1962) pooled data together from the aforementioned Salt and Laing (193 5) study and other early

experiments to divide Trichogramnma egg parasitism behavior into four components. First, the

wasp moves randomly on the substrate (i.e., host plant) and is visually attracted by small obj ects,

which appear to rise above the surface. Eggs must be spotted from a distance of no more than

4mm, and thus walking is a key part of host searching behavior for these wasps (Wajnberg

1994). Plant structure plays a significantly large role in parasitism rates by Trichogramnma

wasps. Factors such as the plant size or surface area, abundance, and diversity of plant parts and

connectivity between individual plants are all keys as to how effective Trichogramnma wasps are

at locating and then parasitizing their hosts (Andow & Prokrym 1990). Second, the wasp climbs

up onto the egg and ritualistically walks all over the surface, making specific movements with its

antennae known as "drumming." Should the egg have already been visited by a different

Trichogramnma female, the wasp will rej ect the egg upon sensing the odor still present on the

chorion of the egg. Otherwise, the parasitoid will proceed across the egg surface. Thirdly, the

parasitoid will begin to drill on the surface of the obj ect, apparently paying little attention to

details of the surface such as texture and smell (provided it doesn't smell like another female

Trichogranmma at which time the parasitoid would arrest its examination and then move on to

another egg) (Klomp & Teerink 1962). These wasps have been observed exhibiting the same

behavior with small seeds and grains of sand, suggesting that texture and smell have little to do

with whether the wasp accepts the egg as a viable host. However, the size of the obj ect will be

assessed, and if development of the wasp offspring is deemed impossible by the adult wasp, the

wasp will immediately cease all further interaction with the egg and move on. Fourth, once the

chorion is perforated by the wasp, she will deposit eggs. Should the host embryo inside the egg








































Figure 4-3. A residence in the area surrounding Greathouse farm, complete with butterfly garden









maximizing intake while also developing strategies to avoid or deter predation (Nylin &

Gotthard 1998). Monarch butterflies (Danaus plexippus) employ a special tactic in deterring

predation. By feeding on Asclepia~s species, they take advantage of the milky latex that is

contained in a sealed system of vessels known as laticifers (Lucansky & Clough 1986). Monarch

larvae are not only able to bypass this plant defense by effectively disabling the latex flow, but

they also are able to sequester and concentrate the cardiac glycosides found in the latex

(Dussourd 1990, Malcolm & Brower 1989). This facet of monarch biology has attracted quite a

bit of attention and consequently has been researched extensively to produce indisputable

evidence that the storage of these cardiac glycosides effectively deters vertebrate predators

(Brower 1969, 1984; Brower & Fink 1985; Glendinning & Brower 1990). However, there are

negative aspects to feeding on plants with latex and cardiac glycosides. In a study by Zalucki et

al. (2001), results suggest that the amount of time spent avoiding excessive latex ingestion

results in slower growth. Because younger and smaller larval stages are more susceptible to

predation by ants, Hemiptera, Neuroptera, and other entomophages, additional time spent in a

vulnerable stage reduces survival of the butterfly larvae (Zalucki et al. 2001). Another negative

aspect of latex is the risk of accidental ingestion. While reports by Rothschild et al. (1977) and

Dixon et al. (1978) indicate that fifth-instar larvae actively search for latex exuded from injured

parts of the plant, Zalucki et al (2001) report that ingestion of latex by first-instar larvae renders

the larvae inactive or cataleptic. This cataleptic state could leave the organism vulnerable to

predation. Although this prediction was not directly tested by them, Zalucki et al. (2001)

observed that catalepsis and disorientation resulting from ingestion of latex by early-instar

butterfly larvae caused larvae to fall off their host plants, increasing the likelihood of being

predated upon by ants and other ground-dwelling predators. In line with these experiments,





H~gNumber 288 997
Location

Figure 3-6. Location had a pronounced effect on egg number, with almost 3 times as many eggs
at Greathouse when compared to Shady Oaks

Table 3-3. Mean % survivor ship among larvae at experimental and control plots for all
geographic locations
Farm Experimental Control
Shady Oaks 43 18
Greathouse 40 10
Florida combined 41 14
El Bosque 83 63


Table 3-4. Mean % survivor ship among larvae at far and near distances in the experimental
plot for all geographic locations
Farm Far Near
Shady Oaks 25 60
Greathouse 25 55
Florida combined 25 58
El Bosque 75 90


Table 3-5. Mean % survivor ship among larvae at far and near distances in the control plot for
all geographic locations
Farm Far Near
Shady Oaks 35 0
Greathouse 0 20
Florida combined 18 10
El Bosque 85 40


200 |-









Each of the farms tested in this experiment report that they have thousands of adult

butterflies present at any one time, as well as thousands of host and nectar plants within the

borders of the farm epicenter. With high levels of butterflies present for breeding at the farm, as

well as the probability of host and nectar plants serving to attract even more butterflies, we can

say that there are high conspecifie densities among butterflies being reared at the farm. In turn,

these high conspecific densities could themselves serve as attractants, according to the

conspecific attraction hypothesis, further increasing levels of conspecifie densities (Ray et al.,

1991). According to Odendaal et al. (1989), butterflies are commonly attracted to areas of high

conspecific densities due to the increased probability of finding a suitable mate. Additionally,

Gilbert and Singer (1973) suggested that a high density of butterflies in an area, especially of the

same species, could be used by a butterfly as an indicator for suitable habitat. The conditions at

the farm in which high densities of butterflies are continually present is a possible force of

attraction that may cause high levels of oviposition closer to the farm epicenter. With the

epicenter providing host plants, nectar plants, and possibly mating opportunities for butterflies, it

is not surprising that there was less oviposition on those plants further away from the farm. This

study, however, did not examine butterfly abundance 0I ithrin the farm epicenter. There is a need

for further research to be done regarding the populations of butterflies actually inside the

butterfly farm epicenter. Although the results from this study indicate that the farm may serve as

an attractive force, the data can only suggest that butterflies are likely to be more abundant closer

to the farm epicenter. Further studies are needed to ascertain whether or not higher butterfly

abundance is caused by attraction, and if so, to what degree.

Larval Predation in Florida and Costa Rica

A very unexpected result of this study was that the presence of the butterfly farm, in all

geographic locations tested, provided for higher levels of larval survivorship. This is true when





































O 2008 John Courtland Whelan













"'~rJ
;4f
r


I


,n~'


~ *I


.i .~
-jg


r' I


' :*~;~~


I -.: -'L' ~Li:i~~-r


Figure 4-7. Blue plastic material used in hopes of stimulating oviposition by M~orpho peleides




























82










OMVENGE, P. I. 2002. The Role of Butterfly Farming in Forest Conservation and Community
Development in Kenya. Swedish University of Agricultural Sciences. Master's Thesis
No. 16

PAK, G. A., I. VAN HALDER, R. LINDEBOOM, AND J. J. G. STREET. 1985. Ovarian egg supply,
female age and plant spacing as factors influencing searching activity in the egg
parasited Trichogramnma sp. Medelelingen van de Faculteit Landb ouwwetenshappen
Rijksuniversiteit Gent 50 (2a) 369-378.

PINTO, J. D., AND R STOUTHAMER. 1994. Systematics of the Trichogrammatidae with emphasis
on Trichogramnma In E. Wajnberg, and S. A. Hassan (Eds.). Biological Control with
Egg Parasitoids, pp. 1-36. Berkshire: CAB International.

PROKOPY, R. J., AND E. D. OWENS. 1983. Visual detection of plants by herbivorous insects.
Annual review of Entomology 28: 40-48.

RABASA, S. G., D. GUITIERREZ, AND A. ESCUERDO. 2005. Egg laying by a butterfly on a
fragmented host plant: a multi-level approach. Ecography 28: 629-639.

RAUSHER, M. D. 1979a. Larval habitat suitability and oviposition preference in three related
butterflies. Ecology 60 (3): 503-511

.1979b. Egg recognition: its advantage to a butterfly. Animal Behavior 27: 1034-1040.

RAY, C., M. GILPIN, AND A. T. SMVITH. 1991. The effect of conspecific attraction on
metapopulation dynamics. Biological Jounral of the Linnaean Society 42: 123-134.

RENWICK, J. A. A., AND C. D. RADKE. 1980. An oviposition deterrent associated with frass from
feeding larvae of the cabbage looper. Trichopulsia ni (Lepidoptera: Noctuidae).
Environmental Entomology 9: 318-320.

AND 1982. Activity of cabbage extracts in tererring oviposition by the
cabbage looper, Trihoplusia ni. Proceedings of the Fifth International Symposium of
Insect-Plant Relationships. Wageningen: Pudoc. pp. 139-43.

ROMEIS, J., D. BABENDREIER, F. L. WACKERS, AND T. G. SHANOWER. 2005. Habitat and plant
specificity of Trichogramnma egg parasitoids underlying mechanisms and
implications. Basic and Applied Ecology 6: 215-236.

ROTHSCHILD, M., AND L. M. SCHOONHOVEN. 1977. Assessment of egg load by Pieris brassicae
(Lepidoptera: Pieridae). Nature 266: 352-355.

SCHLAEPFER, M. A., M. C. RUNGE, AND P. W. SHERMAN. 2002. Ecological and evolutionary
traps. Trends in Ecology and Evolution. 17 (10): 474-480.









These plants were low to the ground, had leaves only 10 cm above the grass line, grew almost

prostrate in the field, and did not have any flowers or other color patterns to distinguish them

from the surrounding grasses. These qualities are nearly the exact opposite of the milkweed used

for the Monarch butterflies. The milkweeds grew vertically, exhibited a height of about V/2 meter,

commonly had flowering parts, and the leaves were well above the grass line.

A second possible reason for the lack of oviposition is an improper choice of host plants,

despite the fact that these were the host plants actively used for rearing of the respective butterfly

species at both farms. Rausher (1979a) suggests that landing frequencies among butterflies may

fluctuate and change seasonally, depending on host abundance and quality. It is possible that

there was a temporary or seasonal host plant shift prior to the start of the experiment. In

addition, Courtney and Forsberg (1988) suggest that landing frequencies may depend on the

relative abundance of a preferred host plant. Should the numbers or quality of these host plants

in the wild have declined recently, or perhaps quantities of another host plant surged prior to the

experiment, landing frequencies, and consequently oviposition, could have been negatively

affected for these butterflies.









arises. The butterfly enclosures could be serving as an ecological trap, in that they suggest

appropriate habitat and substantial mate availability when in fact the opposite is the case.

Ecological Traps

In choosing an appropriate habitat, whether it is to breed, forage, lay eggs, or simply for

refuge, organisms rely on cues exhibited by the environment as indicators. An organism's

decision to choose a particular habitat is largely the result of evolution and an adaptation to rely

on cues that previously have correlated with survival and reproductive success (Williams &

Nichols 1984). With enough evolutionary time and a strong enough correlation, a sudden change

in the environment could endanger the organism by altering the "time-tested" association

between a particular cue (or set of cues) and a predictable outcome for the organism (Levins

1968).

Gates and Gysel (1978) were among the first to use the term "ecological trap" to describe

the increasingly common occurrence of what was described in the previous paragraph. They

observed a situation in which a bird's choice of nesting habitat led to nest failure due to recent

anthropogenic change in the environment. This nest failure broke the connection between the

normal cue and the habitat quality to which these birds were evolutionarily accustomed, and this

resulted in a complete loss of brood by the nesting adults. Susceptibility to ecological traps is

not restricted to birds, as these traps may victimize any organism whose ancestors recognized

and acted upon a tight correlation between a cue and the present or future state of the

environment (clearly an intelligent and historically successful behavior!).

Butterfly farms may present an example of how an ecological trap could work by

displaying "false cues" for habitat quality. Ubiquitous at butterfly farms are specific host and

nectar sources for the butterflies they breed. Similarly, the butterflies bred at farms are typically

the same species that occur in the local area, due to climate, restrictions stated in government





PAGE 1

1 BUTTERFLY ABUNDANCES, LARVAL PRE DATION, AND EGG PARASITISM AS DETERMINED BY PROXIMITY TO BUTTERFLY FARMS IN FLORIDA AND COSTA RICA By JOHN COURTLAND WHELAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 John Courtland Whelan

PAGE 3

3 To my Mom, who has been a great role model and friend

PAGE 4

4 ACKNOWLEDGMENTS First, I would like to thank m y Mom, who has always been there to offer good advice, encouragement and support. I am forever gr ateful for her guidance and dedication throughout my life. I would also like to thank Fra nk and Kim Powell for greatly cont ributing to my interest in science and ecotourism. The oppor tunities they gave me provide d a spark that quickly ignited into a great passion. My entire committee has been of so much help during my graduate education, and I cannot thank them enough for providing me the opportunity to learn from each of them. Thanks go to all those who have advised me along the way, including Tom Emmel, Jaret Daniels, Jackie Miller, John Capinera, and Taylor Stein. Edith and Steven Smith of Shady Oaks, Za ne Greathouse of Greathouse Butterfly Farm, and Ernesto Rodriguez and John F azzini of El Bosque Nuevo were an essential part of my research, as they provided me with help, guidance, and resources that were critical to the success of my study. Their partic ipation, cooperation, and knowledge is greatly appreciated. Lastly, I would like to thank the Chair of my committee and major advisor, Tom Emmel, who has provided me with incredible opportunities that have greatly contri buted to my education and have given me tremendous experiences in the world of ecotourism entomology. He is a great mentor and friend, and a most sin cere thank you is certainly in order.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT...................................................................................................................................10 CHAP TER 1 INTRODUCTION..................................................................................................................12 Butterfly Farming and Ecotourism......................................................................................... 12 Objectives...............................................................................................................................14 Larval Predation............................................................................................................... .......16 Butterfly Parasitism................................................................................................................19 Parasitoid Behavior......................................................................................................... 20 Parasitoids and the Farms................................................................................................ 22 Dispersal.................................................................................................................................23 Factors Affecting Butterfly Dispersal............................................................................. 24 Conspecific density.................................................................................................. 25 Negatively density-dependent dispersal................................................................... 25 Positively density-dep en dent dispersal.................................................................... 26 Dispersal and the Farms.................................................................................................. 27 Ecological Traps............................................................................................................... ......28 Oviposition.............................................................................................................................30 Variation in Egg and Clutch Size.................................................................................... 30 Oviposition Preference.................................................................................................... 31 2 MATERIALS AND METHODS........................................................................................... 34 Oviposition as Index to Butterfly Abundance........................................................................ 34 Egg Parasitism........................................................................................................................36 Larval Predation............................................................................................................... .......36 3 RESULTS...............................................................................................................................47 Main Experiment Butterfly Ovi position in North Central F lorida...................................... 47 Main Experiment: Butterfly Oviposition in Costa Rica......................................................... 50 Larval Experiment: Larval Survivorship as Determined by Distance from Farm Epicenters in North Central Florida .................................................................................... 50 Larval Experiment: Larval Survivorship as Determined by Distance from Farm Epicenters in Costa Rica ..................................................................................................... 51 Egg Parasitism........................................................................................................................51

PAGE 6

6 4 DISCUSSION.........................................................................................................................59 Butterfly Abundance and Oviposition at North Central Florida Farm s.................................59 Main Effect 1: Day..........................................................................................................62 Main Effect 2: Geographic Location............................................................................... 64 Main Effect 3: Transect Station....................................................................................... 67 Larval Predation in Florida and Costa Rica............................................................................ 68 Egg Parasitism in Florida and Costa Rica.............................................................................. 70 Butterfly Abundance and Oviposition at the Northern Costa Rica Farm...............................72 5 CONCLUSIONS.................................................................................................................... 83 LIST OF REFERENCES...............................................................................................................85 BIOGRAPHICAL SKETCH.........................................................................................................93

PAGE 7

7 LIST OF TABLES Table page 2-1 List of butterfly and host plants used................................................................................. 39 3-1 Summary statistics for North cen tral Florida oviposition experim ent............................... 53 3-2 Significance values obtained from the SAS proc mixed procedure ( =0.1) ..................... 53 3-3 Mean % survivorship among larvae at experim ental and control plots for all geographic locations..........................................................................................................56 3-4 Mean % survivorship among larvae at far a nd near distances in th e experim ental plot for all geographic locations................................................................................................56 3-5 Mean % survivorship among larvae at far and near distances in the control plot for all geographic locations .....................................................................................................56

PAGE 8

8 LIST OF FIGURES Figure page 2-1 Entrances to the thr ee butterfly farm s used throughout the experiment............................ 39 2-2 Diagram of farm epicenter (green circle ) with radiation rows of hostplants at 10 m eter transects (yellow dots)............................................................................................. 40 2-3 Dense jungle causing one row of host plants to be shifted slightly ................................... 41 2-4 Several plants needed to be placed cl ose to road s in the Florida experiments.................. 42 2-5 Screened vials and stereoscope used to obs erve the em ergence of parasites or larvae..... 43 2-6 Diagram of larval predation experiment............................................................................ 43 2-7 Controlled environment used to accl im ate Monarch larvae to their host plant................. 44 2-8 Morpho peleides host plant Leguminosa cristobal before and after predator screen application .................................................................................................................... ......45 2-9 Close-up view of predator screen......................................................................................45 3-1 Summary graph of average egg number for each transect ................................................. 53 3-3 Mean egg number at each transect for Shady Oaks Butterfly Farm.................................. 54 3-4 Location and day affected egg number more than just th e sum of their individual roles....................................................................................................................................55 3-5 The day effect on egg number was relatively bell-shaped, with a peak in egg counts for days 4 through 6 after initial placem ent of the potted plants....................................... 55 3-6 Location had a pronounced effect on egg number, with almost 3 times as many eggs at Greathouse when compared to Shady Oaks ...................................................................56 3-7 Mean % survivorship among larvae at experim ental and control plots for north central Florida farms..........................................................................................................57 3-8 Mean % survivorship among larvae at far a nd near distances in th e experim ental plot for north central Florida farms...........................................................................................57 3-9 Mean % survivorship among larvae at far and near distances in the control plot for the northern Costa Rica farm ............................................................................................. 58 4-1 Although most host plants were situated in fields like this at S hady Oaks Butterfly Farm, some plants were in more fore sted areas as seen in the background....................... 76

PAGE 9

9 4-2 More open layout of the area surrounding Greathouse farm ............................................. 77 4-3 A residence in the area surrounding Greathous e far m, complete with butterfly garden... 78 4-4 Large screen enclosure housi ng hundreds of m ilkweed plants..........................................79 4-5 Student groups participati ng in ecotourism may be percei ved as threatening by larger butterfly predators, such as birds and lizards..................................................................... 80 4-6 Heliconia plants around the exterior of the fa rm epicenter could have acted as competition for oviposition in the area.............................................................................. 81 4-7 Blue plastic material used in hopes of stim ulating oviposition by Morpho peleides ........82

PAGE 10

10 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science BUTTERFLY ABUNDANCES, LARVAL PRE DATION, AND EGG PARASITISM AS DETERMINED BY PROXIMITY TO BUTTERFLY FARMS IN FLORIDA AND COSTA RICA By John Courtland Whelan August 2008 Chair: Thomas C. Emmel Cochair: Jaret C. Daniels Major: Entomology and Nematology Butterfly farming is an emerging and rapi dly growing industry, both in developed and developing nations around the worl d. As an industry focused on breeding butterflies for sale, while providing a number of incentives to local communities such as jobs, revenue, education, and conservation, these farms are becoming popular enterprises and ecot ourism destinations. However, little is known about the impacts of these farms from a biological and ecological viewpoint. This study aimed to collect data on 1) wild butterfly abundance, as indicated by oviposition rates, 2) larv al predation, and 3) e gg parasitism rates to de termine any effect that proximity to the farm epicenter might have on each factor. Additionally, the experiment was duplicated in north central Florida, U.S.A and Guanacaste, Costa Rica, to test for any similarities or differences in the aforementione d factors and variables that may be attributable to location. It was found that butterflies tended to oviposit in greater numbers clos er to the farm epicenter in north central Florida, suggesting that the farm coul d serve as an attractive force for gravid female butterflies. Although untested, comp onents of the farm, such as enhanced host and nectar plant biomasses and high levels of conspecific dens ity, could contribute to the higher butterfly abundance closer to the farm. The effect these components have on butte rfly abundances and the

PAGE 11

11 possibility of butterfly attraction to specific ar eas of the farm necessita tes further research. Unexpectedly, there was no ovipositi on recorded during the entire Cost a Rica trial. It is possible that visibility of the host plants and seasonal shifts in host plant preference by the gravid females severely affected ovipos ition in the tropical forest habitat. Larval predation was found to be more frequent in areas distal to the farm epi centers, both 100 meters from the farm as well as several kilometers. Lastly, in all trial sites, no egg parasitism was observed. These results suggest that while gravid female butterflies are more abundant closer to butterfly farms, there is less threat of predation to their offspring when ovipositing near the farm. The initial prediction that the farm may act as an eco logical trap seems unlikely, as, in fact, larval survivorship is significantly higher near butterfly farms.

PAGE 12

12 CHAPTER 1 INTRODUCTION Butterfly Farming and Ecotourism The breeding of butterflies for sale, known as butterfly farm ing, is increasingly regarded as a profitable m icroenterprise busin ess. This is partly due to th e fact that the overall use and the diversity of applications for these alluring le pidopterans are expanding. Today, butterflies are utilized in many ways, ranging from live butterfly releases at weddi ngs to being displayed in live butterfly exhibits in countrie s around the world (Young 1986). In addition to this increasing popularity of these insects, techniques for farming, breeding, and re aring Lepidoptera are becoming more fine-tuned, therefore expanding output and ultimately escalating the success of the individual farmer. Organizations such as the International Butterf ly Breeders Association (IBBA), founded in 1998, are contributing vast am ounts of information to the overall scene. The IBBA coordinates workshops, manuals, and tutorials, allowing amateurs with little to no prior knowledge to initiate breeding operations and rapidl y become proficient at farming butterflies. The resulting accessibility to butterfly farmi ng through a virtually endless supply of knowledge and support, combined with a relatively high demand for these insects, make this industry increasingly more profitable a nd attractive. With some individual farms seeing annual revenues in the 6-digit range, it is easy to see it beco mes desirable to enter this multi-million dollar industry. At the same time, butterfly farming presents a novel situation in which to promote ecotourism. Defined as responsible travel to na tural areas in which to conserve the environment and promote the welfare of local people, ecotourism is intrinsically associated with conservation (TIES 1998). Ecotourism is a powerfu l tool to promote conservation, both directly through investment in conservation programs as well as indirectly through education and

PAGE 13

13 increased awareness, which are key compone nts in many commercial butterfly farming operations. These attractive qualities of butterfly farmi ng render it very appealing to businessmen and naturalists alike. However, an additional sector is now participatin g: rural communities in developing nations. From a conserva tion perspective, this sector is very intriguing. As stated by Omenge (2002), butterfly farming operations in developing nations are contributing to local awareness and conservation of biodiversity, which is a very positive result. Although butterfly farms are typically regarded as positive influe nces in the environment on a macro-scale, one must not overlook the fact that they may influence their environment on a more local scale. Because of the rapidly increasing numbers of farms and the close association that these farms have with ecotourism, it is timely to assess some of the ecological processes, ecological impacts, and general influences that they ex ert on their surrounding ecosystems. A neoteric study by Gordon and Ayiemba (2003) incorporated the monitoring of butterfly population abundances in the forest surrounding butterf ly farms in Kenya in an effort to examine the impact such farms have on the relative butterf ly abundances in the area. After establishing a highly successful butterfly farming operation aimed at deterring local communities from destroying the nearby forest, the Kipepeo project was born. Before the project began, however, butterfly abundance counts were taken. Then, af ter 4 years of farming operations (and after $400,000 of farming revenue had come to the lo cal community!), butterfly abundance counts were taken again (Gordon and Ayiemba 2003). In both instances, these counts were taken by researchers walking transects and counting sighti ngs in each transect, which provided data on 60 species of butterflies. Of these 60 species, 23 were actively being re ared at the farm at the time, while 37 were not. Results showed that but terfly abundances were significantly higher after the

PAGE 14

14 farm had been in operation for the four-year period, but relative abundance rankings were not significantly different among the species surv eyed (Gordon and Ayiemba 2003). In summary, butterfly populations increas ed proportionately during the first four years that the farm was in place. Gordon and Ayiemba concluded that this was not harmful to the natural butterfly populations since the original b alance of common and rarer butterfly species in the local fauna remained intact (Gordon and Ayiemba 2003). Parasitism and predation at butterfly farms have also been investigated in previous studies. Omenge (2002) reported that in smallscale butterfly farms in Kenya, most farmers experienced noticeable predation and parasitism from ants, spiders, wasps, and lizards (Omenge 2002). Not surprisingly, Omenge (2002) found that farmers with more sophisticated butterfly flight cages experienced fewer incidences of parasitism and predation than farmers with makeshift, more rudimentary cages He reasoned that this was due to the disintegration of the screening on the poorly made a nd maintained flight cages, allo wing parasites and predators to have direct access to their host or prey (Omenge 2002). Objectives To further the study of butterflie s in relation to butterfly fa rms, the present project was initiated to investigate three major aspects of butterfly life history, all with respect to the butterflies proximity to the central operational ar ea of the butterfly farm, which will be referred to hereafter as the farm epicen ter. The first goal was to an alyze butterfly abundance based on oviposition. Contrary to previ ous studies, such as Gordon and Ayiemba (2003), which used visual sightings of adults as indicators for a bundances, my study used the deposition of eggs as an indicator for butterfly abundance. While e gg censusing will not measure the presence of butterflies not laying eggs, such as males or non-gr avid females, this method allows continuous data collection over 24-hour period s, as new egg counts were reco rded regularly every 24 hours.

PAGE 15

15 It was hypothesized that proximity to the farm would result in higher bu tterfly abundances so that the highest abundances would be observe d right next to the farm epicenter. The second research goal was to investigate la rval predation rates based on proximity to the farm epicenter. It was hypothesized that la rval predation rates woul d be highest at points closest to the farm. The third and final goal was to study the rate of egg parasi tism, again based on proximity to the farm epicenter. It was hypothesized that close proximity to the farm would yield higher rates of egg parasitism. Application of this studys re sults could serve to improve the ecological health of the farm ecosystem, its surrounding area, and possibl y result in the enhancement of overall productivity. Butterfly farming operations, both do mestically here in the U.S. as well as in foreign countries, are often situ ated in relatively remote rural areas with significant amounts of surrounding undeveloped land. Relative to urban or suburban settings, these farms typically occupy a more pristine setting in which the surrou nding ecosystem is healthier with higher levels of biodiversity. Although this is not always the case, as some farms are indeed proximal to urban areas, it is of great conser vation concern to better understand th e habits of all life stages of butterflies living within proximity to butterfly farms in order to ensure that these enterprises continue to contribute to bi odiversity conservation and the principals of ecotourism. Such a study is also beneficial to the butterf ly farm as a way to in crease understanding of how butterflies are parasitize d, preyed upon and their ovipos ition habits. By quantifying parasitism rates, the farming operation may be able to adjust management practices in order to ensure higher productivity through lowering parasiti sm rates of the species actively being bred. This higher productivity, at a human social and community level, could result in more revenue

PAGE 16

16 and job opportunities to the co mmunity involved. By measur ing butterfly abundance rates, scientists are able to better understand the ecological dyn amics and influences of a butterfly farm and its impact on the surrounding areas. Should i rregular patterns of butterfly distribution and abundance occur (i.e., butterflies fo und to be not evenly disperse d throughout the farm property), it was the objective of this study to not only observe and record such patterns, but also to provide interpretation of the factors in fluencing such irregularity. More importantly, some investigators have suggested that the farm may act as an ecological trap, where th e butterflies are being attracted to the farm epicenter under the auspic es of reward for homing in on an environmental cue (Schlaepfer et al 2002). Ecological traps will be discussed further in a later section starting on page 29. Finally, at a conservation level, under standing the reasons for patterns of parasitism, predation and butterfly abunda nce in areas surrounding the farm will allow for better management of the farm. Ultimately the hope w ould be that the results of this study might provide supportive evidence that will minimize any potential negative influence the farm may have on its environment, and result in a more stabilized, productive, a nd healthy ecosystem. This, in turn, will produce a healthier farm ecosy stem and likely increase the sustainability and productivity of the butterfly farm and th e surrounding human communities that may be economically dependent on the farm. The remainder of this introduction will detail the major topics addressed in this study, including larval predation, egg parasitism, oviposition, disp ersal, and ecological traps. Larval Predation The fitness of most organisms is closely related to the organisms age and size at sexual maturity (Stearns 1992). Reaching an optimal combination of age and size as an adult is largely governed by how an organism performs in its juve nile stages, particularly concerning growth rate (Stearns 1992). Regarding juvenile survivorship, one can see the benefits of a rapid growth rate

PAGE 17

17 (provided that there is sufficien t food supply available). Theoretically, a faster growth rate results in less time spent in the vulnerable juvenile period, reducing the amount of time the larvae is exposed to threats, and increasing the probability that the immature organism will survive until reaching sexual maturity (Stearns 1992). However, Arendt (1997) suggests that oftentimes organisms grow at a slower rate than they are physiologically capa ble. If it is indeed true that a shorter juvenile pe riod would lead to greater overall success of the animal (i.e., less time susceptible to predation, para sitism, and other threats), why wouldnt the animal attempt to maximize its growth rate? The answ er is relatively simple. Grow th rate and predation level are both functions of activity level in many or ganisms (Werner & Anholt 1993). Commonly, organisms that exhibit a maximum growth rate also require maximum nutrition and, therefore, devote more time to foraging activity, which put s the animal out in the open more often and increases its susceptibility to pred ation. Consequently, larval growth patterns are likely to be the result of making strategic decisions where th e organism weighs the costs and benefits of various growth trajectories and chooses an optimal rate dependent upon its particular environment (Abrams et al 1996, Nylin & Gotthard 1998, Gotthard 1999). In a laboratory experiment, Gothard (2000) found that there is indeed a trade-off between butterfly larvae growth rate and predation risk, using the butterfly Pararge aegeria (Nymphalidae: Satyrinae). When comparing a fast-growing larvae and slow-growing larva e, they found that the daily predation risk was 30 % higher in their faste r-growing larvae treatme nt (Gothard 2000). Therefore, we see that larval predation may be more of a problem for species that exhibit naturally high growth rates. Butterfly larvae are at particular risk to pred ation due to their feed ing habits. They are essentially feeding machines that devote their en tire juvenile lives, as well as body plans, to

PAGE 18

18 maximizing intake while also developing strate gies to avoid or dete r predation (Nylin & Gotthard 1998). Monarch butterflies (Danaus plexippus ) employ a special tactic in deterring predation. By feeding on Asclepias species, they take advantage of the milky latex that is contained in a sealed system of vessels known as laticifers (Lucans ky & Clough 1986). Monarch larvae are not only able to bypass this plant defe nse by effectively disabling the latex flow, but they also are able to sequester and concentr ate the cardiac glycosides found in the latex (Dussourd 1990, Malcolm & Brower 1989). This f acet of monarch biology has attracted quite a bit of attention and consequently has been re searched extensively to produce indisputable evidence that the storage of these cardiac glycos ides effectively deters vertebrate predators (Brower 1969, 1984; Brower & Fink 1985; Glendi nning & Brower 1990). However, there are negative aspects to feed ing on plants with late x and cardiac glycosides. In a study by Zalucki et al (2001), results suggest that the amount of time spent avoidi ng excessive latex ingestion results in slower growth. Because younger and sm aller larval stages are more susceptible to predation by ants, Hemiptera, Neuroptera, and other entomophages, additional time spent in a vulnerable stage reduces survival of the butterfly larvae (Zalucki et al 2001). Another negative aspect of latex is the risk of accidental ingestion. While reports by Rothschild et al (1977) and Dixon et al (1978) indicate that fifth-in star larvae actively search for latex exuded from injured parts of the plant, Zalucki et al (2001) report that ingestion of latex by first-instar larvae renders the larvae inactive or cataleptic. This catalepti c state could leave the organism vulnerable to predation. Although this prediction was not directly tested by them, Zalucki et al (2001) observed that catalepsis and diso rientation resulting from inges tion of latex by early-instar butterfly larvae caused larvae to fall off their host plants, increasing the likelihood of being predated upon by ants and other ground-dwelling predators. In line with these experiments,

PAGE 19

19 many other studies of Lepidoptera have found th at low survivorship of early instars is responsible for most of each generations mortal ity, a truly grave realiz ation (Cohen & Brower 1982, Zalucki & Kitching 1982, Dempster 1983, Kyi et al 1991). For these re asons, third instar larvae were the subject for this part of the st udy. Third instars would not be as vulnerable to catalepsies, while also having less of a proclivity to sequester hi gh levels of cardiac glycosides. Despite these studies, there are very few published studies that estimate natural predation risks for butterfly larvae. However, one of th ese few studies found staggering results. Bernays (1997) made a comparison of predation risks whil e Lepidoptera larvae were feeding versus when they were resting. He found that predation le vels while feeding increased three-fold for Manduca sexta and 100-fold for Uresiphita reversalis. There is a great need for more research to be done examining the predation risk for but terfly larvae in a natural environment. Due to the extensive literature base, as well as their prevalence in butterfly farm cultures and in the Florida wi ld, Monarch butterflies (Danaus plexippus) were selected as the principal study organism for my experiment at north cen tral Florida butterfly farms. Due to the predominant presence and abundance of the Common Morpho (Morpho peleides) at Costa Rica butterfly farms, this Morpho was selected as th e primary study organism at Costa Rica butterfly farms. Butterfly Parasitism The relative importance of butterfly parasitism studies is being realized more and more as butterfly captive breeding and ecotourism programs continue to emerge and grow. While the majority of past studies have focused on the positiv e effects, or at least the theoretically positive effects, of parasites and parasito ids of Lepidoptera that are used in biocontrol, the research and findings on these situations are relevant a nd contribute to the overall understanding of lepidopteran parasitoid behavior. Nevertheless, th ere is a need for additional research to be done

PAGE 20

20 on the possible negative effects of lepidopteran para sites, particularly those parasitoids affecting butterfly farming, captive breeding, and butterfly ecotourism programs. When dealing with parasitoid s that affect insects in th e order Lepidopter a, butterfly farmers and biologists alike can identi fy several types of insects that could pose potential problems, such as Tachinid flies (Diptera: Tachinidae) and Braconid wasps (Hymenoptera: Bradonidae), there is one key wasp group that overwhelmingly presents the greatest threat: species of the egg parasitoid, genus Trichogramma (Hymenoptera: Trichogrammatidae). Many species in this genus are impor tant natural enemies and are us ed in inundative and inoculative biocontrol programs throughout the world (Li 19 94, Smith 1996). While many species in the genus Trichogramma are generalists, the better-known species are sp ecialists on parasitizing eggs of the members of the order Lepidoptera (Pinto & Stouthamer 1994). Because of their extremely small size, ranging from 0.2 to 1.5mm, these tiny wasps present a much greater problem than other parasitoids for the captive breeding of Lepidoptera (Pinto & Stouthamer 1994). This is because exclusionary tactics such as screens, netting, and cages need to be almost perfect (i.e., no gaps, tears, etc. ) to ensure that these nearly microscopic insects do not have access to breeding centers and the Lepidoptera eggs within. While newer exclusionary technology allows for better parasite control, the costs of these more effective methods are often prohibitive to butterfly farmer s in developing nations (Gordon & Ayiemba 2003). Because butterfly farming and ecotourism in developing nations are increasingly becoming vital sources of revenue and a means for biodiversity conserva tion, a better understanding of egg parasitoid behavior in these areas is critical. Parasitoid Behavior The behavior of Trichogramma parasitoids was studied as early as 1935 by G. Salt and J. Laing. Extensive experim ents and reviews ha ve been published subsequently, which have

PAGE 21

21 contributed critical knowledge to our understanding of these organisms. Klomp and Teerink (1962) pooled data together from the aforementi oned Salt and Laing (1935) study and other early experiments to divide Trichogramma egg parasitism behavior into four components. First, the wasp moves randomly on the substrate (i.e., host plan t) and is visually attr acted by small objects, which appear to rise above the surface. Eggs mu st be spotted from a distance of no more than 4mm, and thus walking is a key part of host searching behavior for these wasps (Wajnberg 1994). Plant structure plays a significan tly large role in parasitism rates by Trichogramma wasps. Factors such as the plan t size or surface area, abundance, and diversity of plant parts and connectivity between individual plants are all keys as to how effective Trichogramma wasps are at locating and then parasitizi ng their hosts (Andow & Prokrym 1990). Second, the wasp climbs up onto the egg and ritualistically walks all over the surface, maki ng specific movements with its antennae known as drumming. Should the egg have already been visited by a different Trichogramma female, the wasp will reject the e gg upon sensing the odor still present on the chorion of the egg. Otherwise, the parasitoid w ill proceed across the egg surface. Thirdly, the parasitoid will begin to drill on the surface of the object, apparent ly paying little attention to details of the surface such as texture and sme ll (provided it doesnt smell like another female Trichogramma, at which time the parasitoid would arrest its examination and then move on to another egg) (Klomp & Teerink 1962). These wasps have been observed exhibiting the same behavior with small seeds and grains of sand, s uggesting that texture and smell have little to do with whether the wasp accepts the egg as a viable host. However, the size of the object will be assessed, and if development of the wasp offspr ing is deemed impossible by the adult wasp, the wasp will immediately cease all furt her interaction with the egg and move on. Fourth, once the chorion is perforated by the wasp, she will deposi t eggs. Should the host embryo inside the egg

PAGE 22

22 be too far developed, or perhaps if it has alrea dy been parasitized, the wasp will abandon the egg and move on (Klomp & Teerink 1962). These wasp parasites can be costly problem s for Lepidoptera, a fact which has deemed them powerful bio-control agents on pest Lepidopt era in agriculture. Unfortunately, they also parasitize non-pest Lepid optera that can thereby cause costly problems for butterfly farmers. As mentioned earlier, exclusionary measures are fre quently used to protect butterfly stock from unacceptably high levels of parasitism. However, these measures are typically employed only in special labs and buildings for butterfly eggs and other immature stages. Trichogramma wasps still present problems in other areas of the bu tterfly farming operation, such as outdoor plant nurseries and gardens. Parasitoids and the Farms A common sight at butterfly farms is host and nectar plants growing in relatively vast numbers around the property. Some of these plants are found under the cover of screens in shadehouses, while others are out in the open to ta ke advantage of sunlight as well as minimizing overhead costs to the farm. In both cases, these pl ants have the potential to attract parasitoids, creating a risk for the farm. As plants in the wi ld serve as resources to butterflies, so do plants that are found in the open areas at butterfly farms a nd breeding centers. The potential for high densities of butterfly eg gs at these host plant nurseries w ould naturally attract parasitoids, especially Trichogramma wasps. Additionally, the close spacing of plants, common in these butterfly farm nurseries, makes it very easy for para sitoids to walk and jump to adjacent plants (Romeis et al 2005). Research suggests th at this augments parasitoid dispersal, which could make these areas prime habitats for Trichogramma wasps and other egg para sitoids (Surverkropp 1997). If this is indeed true, it means that bu tterfly farms could be supporting high populations of parasitoids in the same genera l area as where they harvest and store butterfly e ggs. There is a

PAGE 23

23 need for research that investigates the rate of egg parasitism within proximity to the butterfly farm to critically examine this problem. Dispersal Generalities and supportive docum entation derived from previ ous dispersal studies have been difficult, mainly due to the amount of variat ion between the definitions of dispersal as well as the great number of differences in spatial scal es used in such studies. Perhaps the two most common definitions for dispersal are (1) movement from the natal patch to the breeding patch, and (2) movement between breeding patches (Clobert et al 2001). However, these definitions do not mention nor take into account the distances or even other reasons for dispersal, such as foraging, for example. To avoid complicated semantic issues, this study will adopt a broad definition of dispersal as stated by Bowler a nd Benton (2005), which defines dispersal as any movement between habitat patches, and habitat pa tches as areas of suitable habitat separated in space from other such areas, regardless of distance. The consequences resulting from dispersa l are great for both the individual actively dispersing as well as the populat ions involved (i.e., the popula tion that the individual is immigrating to as well as the population that th e individual is emigrati ng from). The act of simply moving from one patch to another may af fect individual fitnes s, population dynamics, and genetics, as well as sp ecies distributions (Dunning et al 1995, Hanski & Gilpin 1997, Hanski 1999, Clobert et al 2001). The well-founded link be tween dispersal and population dynamics necessitates a better understanding of the patterns, causes a nd consequences of dispersal. This is critical to our overall unde rstanding of population management and our ability to predict population responses to changes in the environmen t (Bowler & Benton 2005).

PAGE 24

24 Factors Affecting Butterfly Dispersal There are a num ber of different factors that in fluence a butterfly to either leave or remain in a patch of habitat. Some f actors are individual characteristics of the organism such as age or sex (Kuussari et al 1996) while others are char acteristics of the habitat, such as the presence of nectar sources and host plant density (Baker 1969, Shreeve 1995). The larval resource (host plants) and adult resource (nectar plants, sap flow s, rotting fruit) are im portant ecological factors affecting movement in butterflies, as well as other holometabolous invertebrates (Brommer & Fred 1999). Similarly, the sex of the butterfly is co rrelated with the rate of dispersal, as is the insects age. However, there is contention am ong experts as to the nature of the correlation between sex and age with ra tes of dispersal (Kuussaari et al 1996, Lawrence 1987). Although these single factors may exhibit a certain degree of correlation with dispersal, one must question whether correlation proves causa tion. As seen in Bowler a nd Benton (2005), many of these single factors are associated with one unifying condition: conspecific density. Conspecific density, which measures the density of organisms of the same species in a given area, is seen to have a great influence on butterfly movement and dispersal (Dethier & MacArthur 1964). While separate fa ctors such as age, sex, and ha bitat resource av ailability may play their individual parts, cons pecific density may be the most powerful, as it lumps many of the possible factors that govern dispersal rates into one central theme. Food availability, sex ratio, and parasite/parasitoid presence are all co rrelated to conspecific density (Bowler & Benton 2005). Therefore, understanding the e ffect that conspecific density ha s on dispersal is useful in understanding the entire system of dispersal. However, the correlation between conspecific density and butterfly dispersal is not as definitive as, say, the correlation between nectar source presence and butterfly movement into or out of a patch. With no surprise, there is a positive correlation between butterfly movement into th e patch and the presen ce of high numbers of

PAGE 25

25 nectar sources (Shreeve 1995). Ne vertheless, understanding the e ffects that conspecific density has on dispersal is key in this study of butterf ly abundances at butterfly farms, which have artificially high conspecific densit ies due to their butterfly stock and/or dense planting of nectar sources inside and outside the cages. Conspecific density Previous studies have observed that individuals disp ers e at a higher rate away from an area when the area exhibits a low conspecific de nsity, suggesting a negativ ely density-dependent dispersal (Gilbert & Singer 1973; Kuussaari et al 1996, 1998), while other observations find that individuals disperse at higher rates away from areas exhibi ting high conspecific density, suggesting a positively density-dependent disp ersal (Dethier & MacArthur 1964; Odendaal et al 1989; Baguette et al 1996, 1998). The next two sections will concentrate on the differences between these two concepts and common ar guments in favor of one or the other. Negatively density-de pendent dispersal Negatively density-dependent di spersal is a term used to describe a situation in which individuals disperse at a higher rate away from an area when the area exhibits a low conspecific density (Gilbert & Singer 1973; Kuussaari et al 1996, 1998). According to Hanski et al (1994) and Kuussaari et al (1998), this could be the result of mate scarcity, as a butterfly inhabiting an area with low conspecific density would have a lesser chance of finding an appropriate mate and may choose to leave the patch in order to hopefull y increase chance of finding a mate elsewhere. The concept of high disperal rates away from area s of low conspecific density gave rise to the conspecific attraction hypothesis, which states that high density popul ations will result in higher immigration rates to and lower emigration rate s away from the habitat in question (Ray et al ., 1991). There are two supporting ar guments to this hypothesis. The first, given by Odendaal et al (1989), is that butterf lies are attracted to and ultimatel y remain in habitats with high

PAGE 26

26 conspecific densities due to highe r probabilities of locating suitable mates there. Specifically, males would likely be attracted to areas of high female density, (Hanski et al 1994, Kuussaari 1998). The second argument was given by Gilbert and Singer (1973), who argued that a high density of butterflies in an area, especially of th e same species, could be an indicator for suitable habitat. Therefore, an area with high consp ecific density would trigger emigration from areas with low conspecific density as new adults would view the high density of same-species as a sign of suitable habitat, perhaps containing ample re sources such as water, nectar, and host plants. Positively density-dependent dispersal On the other hand, there are also argume nts supporting positively density-dependent dispersal, which is when indivi duals disperse at higher rates aw ay from areas exhibiting high conspecific density (Dethier & MacArthur 1964; Odendaal et al 1989; Baguette et al 1996, 1998). These supporting arguments most often have to do with social interactions within the patch (Shapiro 1970; Odendaal et al 1985, 1989; Baguette et al 1996, 1998). For instance, in some species of butterflies, males will vigorously pursue females in order to mate with them. While this may be effective for males that are pursuing un-mated females, sometimes the females have already mated and are reluctant to mate again. This does not often dissuade the males, as studies have observed them harassing females to the point of the female leaving the patch (Odendaal et al 1985). Should this happen frequently and repeatedly, a likely scenario in a high density population, the re sult would be a high dispersal rate away from the area by females. This, in turn, could trigger a similar dispersal pattern among males that may then leave the area too, in pursuit of females or in s earch of an area with more females (Baguette et al 1998). A high conspecific density in a particular patch may also directly affect habitat quality of the patch, which could lead to higher emigra tion rates should the patc hs density surpass the

PAGE 27

27 patchs carrying capacity (Dethier and MacArthur 1964). As the butterflies deplete the host and nectar plant resources in the habitat, a migra tory threshold would be reached (Baker 1984). This threshold is a theoretical limit to the rate of migration so that as the threshold is exceeded, dispersal away from the patch would begin to su rpass dispersal into the patch, continuing this way until returning to a stable density. This stable density would be below the carrying capacity and allow key resources of the habitat such as n ectar and host plants to regenerate, allowing for more immigration in the future. Dispersal and the Farms The notion of conspecific density is of par ticular importance to this study of butterfly farms, which have artificially high conspecific dens ities. It is clear that farm enclosures are artificially stocked with butterflies, yielding this artificially high conspe cific density. However, it has not been documented whethe r or not butterflies occur in high numbers around the exterior of the farm epicenter. The enclosures at butterfly farms present a special circumstance when considering habitat patches and conspecific densities. They may be treated as separate hab itats due to the walls excluding immigration and migration. At the same time, however, they may still serve as an attractive force (i.e., presence of host plants, ne ctar sources, pheromones and visual signals emanating from potential mates, et c.) to wild populations of butterflies, in which case they could be considered part of a larger patch that includes all areas of the butte rfly farming operation. According to Hanski et al (1994) and Kuussaari et al (1998), butterflies may be attracted to high conspecific densities in search of an a ppropriate mate. Additionally, Gilbert and Singer (1973) suggested that high conspeci fic densities may also attract bu tterflies when acting as a cue for good habitat quality. Should these theories ap ply to butterfly farms as well, a great problem

PAGE 28

28 arises. The butterfly enclosures could be serving as an ecological trap, in that they suggest appropriate habitat and substantial mate availabi lity when in fact th e opposite is the case. Ecological Traps In choosing an appropriate habitat, whether it is to breed, forage, lay eggs, or simply for refuge, organisms rely on cues exhibited by the environment as indicators. An organisms decision to choose a particular habi tat is largely the result of evolution and an adaptation to rely on cues that previously have correlated w ith survival and reproduc tive success (Williams & Nichols 1984). With enough evolutionary time and a strong enough correl ation, a sudden change in the environment could endanger the organi sm by altering the time-tested association between a particular cue (or set of cues) and a predictable outcome for the organism (Levins 1968). Gates and Gysel (1978) were among the first to use the term ecological trap to describe the increasingly common occurren ce of what was described in th e previous paragraph. They observed a situation in which a birds choice of ne sting habitat led to nest failure due to recent anthropogenic change in the environment. This nest failure broke th e connection between the normal cue and the habitat quality to which these birds were evolutionarily accustomed, and this resulted in a complete loss of brood by the nesting adults. Susceptibility to ecological traps is not restricted to birds, as th ese traps may victimize any orga nism whose ancestors recognized and acted upon a tight correlation between a cu e and the present or future state of the environment (clearly an intelligent and historically successful behavior!). Butterfly farms may present an example of how an ecological trap could work by displaying false cues for habitat quality. Ubiqu itous at butterfly farms are specific host and nectar sources for the butterflies th ey breed. Similarly, the butterfl ies bred at farms are typically the same species that occur in the local area, due to climate, restrictions stated in government

PAGE 29

29 permits, etc. Additionally, these plants are often kept outside in order to maximize sunlight and minimize overhead cost. Combining these three va riables, there is a high chance that native butterfly populations in the area will find the farm suitable as an area in which to oviposit on the host plants, or to feed on the nectar sources. Although this situation seems lik e it might be no more harmfu l than a butterfly garden, there are several key differences Firstly, these plants are fr equently moved, pruned, and even de-egged (eggs are removed just before using them in the butterfly enclosures to safeguard against the introduction of egg para sites in quarantined zones). These disturbances could cause the eggs to fall off or be removed from the pl ant. The second way that these nurseries differ from a butterfly garden is that the host plants are often very de nse and clustered together in a homogenous pattern. This enables parasites to easily walk and jump to different plants (Romeis et al 2005). Butterfly gardens ar e typically more heterogenous in the way the plants are arranged, serving as natural defenses to parasitoids. Lastly, predat ors such as vespid wasps, ants, lizards and birds are likely to exploit a predic table source of food in th ese nurseries, predating upon adults and larvae. It is clear that if th e conditions are right, nati ve butterfly populations could be in danger of falling into an ecological trap. From an environmental standpoint, the presen ce of an ecological trap could be very detrimental, and the problem requires further investigation. However, it is not just an environmental problem; the farm could also be ne gatively impacted as well. In the case of both parasitoids and predators, high levels could pose problems for the farms butterfly stock. Predators and parasitoids may attempt to enter breeding and rearing cent ers, causing significant economic loss to the farm.

PAGE 30

30 It was the goal of this study to conduct a su rvey of butterfly abunda nces in relation to their proximity to the butterfly farm. Should w ild butterflies be significantly more abundant in areas closest to the farm, this may suggest that they are indeed being attracted to the farm. However, additional studies are necessary to either dismiss this as insignificant or to verify this attraction and distinguish this f actor from the possibility that escaped larvae or adults are artificially inflating populati on numbers on the outskirts of the farm. In addition, larval predation and egg parasitism also need to be exam ined in this study as a function of proximity to the butterfly farm in an effort to docum ent all facets of a possible ecological trap. Oviposition The evolutionary process of insect egg laying, or oviposition, is a topi c that has received a significant am ount of attention in past studies. As with other insects, oviposition in butterflies has been studied from many different angles. T hose pertaining to variations in egg and clutch size, oviposition preference and oviposition habita t have particular appl ication to this study. Variation in Egg and Clutch Size Egg size is a valuable measurement when studying insects, as it is closely related to fitness (Fox & Czesak 2000, Fischer et al 2003). In an ultimate sens e, it involves how mothers choose to allocate resources among their offsprin g. Almost always, as energy available for reproduction is held relatively constant, a larger egg size is correlated with a smaller clutch size, and vise versa (Smith & Fretwell 197 4). However, it is not clearly known how a clutch is optimized by selection for either numbers, or for individual egg size. Although it may not be a panacean explanation for why any gravid fema le insect would choose greater egg size over greater clutch size, one study proposed that there is a correlat ion between temperature and egg size (Atkinson 1994). Atkinson (1994) coined the term temperature-size rule which essentially states that as the temperature experienced by the mother is lowered, the si ze of her eggs become

PAGE 31

31 larger, while the total number of eggs she lays is lessened. These initial beliefs have been confirmed in many recent studies (Azevedo et al 1996; Yampolski & Scheiner 1996; Atkinson et al 2001; Fischer et al 2003, 2006a, 2006b; Steingenga & Fisc her 2007) and could provide insight into butterfly emergence and dispersal patterns. This plasticity between egg size and clutch size is not an even tradeoff, however, as total reproductive investment must be increased at higher temperatures (Avelar 1993 ; Ernsting & Isaaks 1997, 2000; Fischer et al 2003). For example, lifetime fecundity wa s almost twice as high at 27 C than at 20 C in a study by Fischer et al (2003). Although the mothers biological response of egg size and clutch size to temperature was gradual (two to three days of cold could alter egg layi ng behavior) in Fischer et al .s study (2003), it was also found to be reversible after two to three days of warm weather. We can surmise from these studies that ovipos ition will be in greater numbers in warmer environments than in colder environments. Oviposition Preference Most butterf lies are oligophagous. This habi t of feeding on host plants from a very limited number of plant families, oftentimes onl y a single family, suggests that there may be basic similarities within plant families that signal or cue a bu tterfly to oviposit (Feeny 1991). The host plant may be attracting th e butterfly by exuding a chemical, visual, or tactile cue. In any case, it is the job of the adult butterfly to choose the best host plant for its soon-to-be offspring. Larvae of many phytophagous and holom etabolous insects are to a certain degree immobile and are not able to search for appropriate food over compara tively large distances (Rausher 1979a). Females that oviposit on plants with low larval survi vorship or poor larval growth will likely leave fewer descendents than females that oviposit on the more suitable host

PAGE 32

32 plants (Rausher 1979a). For this reason, it is imperative to the su rvival of the species that the female chooses an appropriate host plant correctly the vast majority of the time. Habitat: Habitat is among the more important in fluences regarding host-plant selection among insects. A habitat that simply contains specific host plants is not enough for that habitat to be preferential to gravid female butterfli es (Rausher 1979a). In fact, many ovipositing butterflies do not use all of the habitats in wh ich the appropriate host pl ants grow (Shapiro & Carde 1970, Ehrlich et al 1975). Egg laying patterns could be the result of patch size, connectivity, and landscape matrix as well (Rabasa et al 2005). Plant-insect interactions within a site may also depend on processes occurring at larger spatial scales (Tscharntke & Brandl 2004). Nevertheless, patch size and landscape connectivity are pos tulated as key determinants of butterfly presence in fragmented landscapes, due to the fact that smaller patches experience higher rates of extinctions (James et al 2003). A study by Rabasa et al (2005) found that the probability of egg presences and oviposition for a rare European butterfly did not differ significantly between large and small patches. Some previous studies suggest that oviposition search behavior continually evolves in order to ensure that females lay eggs in the habitats that are mo st well suited for the growth and development of the juvenile stages (Gilber t & Singer 1975, Wiklund 1977). As reported by Hovanitz and Chang (1963), food plant choice a ppears to be at least partly genetically determined in many insect species, with Lepidop tera being no exception. Therefore, one could expect that differences in the suitability of food plants to infl uence the evolution of genes or alleles controlling oviposition behavior (Rausher 1979a). Forister (2004) tells us, however, that there is a positive correlation between female preference of optimal host plant and larval pe rformance and this has been found and cited in

PAGE 33

33 studies of various insect families (Mayhew 1997, Harris et al 2001, Craig & Ohgushi 2002). Nevertheless, this presumption of a strongly po sitive correlation has been integrated in many biological models and taken as fact in many instances. Also termed the nave adaptationist hypothesis, by Courtney and Kibota (1990), this is the most common hypothesis in regards to the selection of host plants by phytopha gous insects. It states that natural selec tion effectively promotes female preference such that eggs are laid on the host plant that is the best host for her offspring (Levins & MacArthur 1969). Reiterati ng what Raucher (1979a) said, there are two ways that the plant can serve as best host: one way resulting in the highest survivorship rate among offspring, and the other being the relative size of the offspr ing without regard to overall survivorship rates (Rausher 1979a). Oviposition is used in this study as an indicator for butterfly abundance. By standardizing the oviposition substrate (i.e., host plant), and environmental conditions such as seasonality and landscape connectivity, any fluctu ations in the number of eggs deposited should indicate a fluctuation in bu tterfly abundance in the area.

PAGE 34

34 CHAPTER 2 MATERIALS AND METHODS Oviposition as Index to Butterfly Abundance This experiment occurred in a total of thr ee geographic locations, with all measurements and experiments set up identically in each location. The first two trials were conducted in north central Florida, at Shady Oaks Butterfly Farm and Greathouse Butterfly Farm. The third trial was conducted in northern Guanacaste Province, Cost a Rica, at El Bosque Nuevo Butterfly Farm (Figure 2-1). For each location in this study, the researcher designated an epicenter, as shown in Figure 2-2. The epicenter containe d all butterfly host plants being grown at the farm (either for sale or for use in rearing and breeding), all butterfly cages, all rearing rooms, laboaratories, and all gardens. The farm epicenter is essentially the operational part of the farm. If there was farm property not specifically used in the butterfly farming operation (such as access roads, parking lots, etc.), it was excluded from the epicenter. The epicenter circumference was then measured using a rolling-wheel electronic tape measurer. A total of nine rows of potted (one-gallon si ze) butterfly host plants were placed, starting at the perimeter of the epicenter and radiating out to a distance of 100 meters for each row. Each 100-meter row was then divided further so that one host plant was placed at each 10-meter interval. Therefore, 11 host plants were pl aced along each row (Figure 2-2). In order to maintain the plants health, a water saucer was placed under each pot to maximize water retention. In order to distinguish the experiment al plants from other nearby conspecifics, orange marking flags were used to identif y the experimental plants. Each host plant row was evenly spaced and angled from the two adjacent rows as accurately as possible so that no two rows were closer than other rows. Using the previous

PAGE 35

35 measurement of circumference, the researcher wa s able to calculate e qual distances and equal angles for the spacing between all rows. A total of three species of host plants were used at each butterfly farm, with each intended to elicit oviposition from a specific specie s of butterfly. See Table 2-1 for the complete list of host plants and butte rflies. Different species of host plan ts were used for a different set of butterflies at the Costa Rica butterfly farm (T able 2-1), but for both north central Florida locations the same plants were used. To recapit ulate, one plant was placed at each 10m transect for each row and there were nine rows of host plan ts total. With three different host plants per farm, each host plant was replicated three times so that there were three rows for each plant. Each row contained 11 plants, from a distance of zero (adjacent to the epicenter perimeter) to a distance of 100 meters. Therefore, 99 indi vidual plants were used for each farm. Challenges in placing the plant rows presented th emselves in the experiment. First, each farm was situated within 100 meters from a roa d. Therefore, several pl ant rows had to cross these roads at which points the researcher preser ved the 10-meter spacing as much as possible. In the case of Costa Rica, dense jungle made 10-meter spacing difficult with two of the nine rows, and as a result the angle between them was s lightly more acute than the other rows in this situation (Figure 2-3). In the case of the north central Florida farms, the plants were placed on the edges of roads (Figure 2-4), which caused several transect intervals to be slig htly longer than 10 meters. Another problem with having roads so close to th e farm being studied was that several plants placed near the road were stolen. However, this happened only once, and the plants were able to be replaced on the following day, thus minimizing the loss of data.

PAGE 36

36 Once the experiment was set up, data were then collected on the number of new eggs found on each plant per day (i.e., every 24 hours). These eggs were subsequently used in the next part of the experiment (see Egg Parasitism section Page 37). All eggs were removed at the time of data collection each day. This was to prevent gravid females from passing up these plants when ovipositing in the following days, due to the possibility of females of that species having responded to past selection to relieve conspecific egg competition by visually inspecting host plants for the presence of pr eviously laid eggs. Data collection was repeated on a daily basis (i.e., every 24 hours) over a 10-day period re sulting in 30 replicat e host plant at each transect number (distance from ep icenter) for each farm location. Egg Parasitism The second part of the experim ent was to test for egg parasitism at varying distances from the butterfly farm epicenter. Using eggs from the previous oviposition experiment, a total of 12 eggs was collected at random from each transect number and retained for use in this eggparasitism study. Eggs were collected from the plants every 24 hours, at 8:00 A.M. (this samples eggs exposed to parasitism incidence from the preceding 24 hours). Afte r collecting these eggs, they were placed individually in plastic containers covered with a fine mesh rubber-banded on top to allow air flow but to prev ent any exit of the hatc hed larvae, or hatched parasite. The mesh also blocked entrance of subsequent searching parasites. Then, the eggs were observed on a daily basis until emergence of larvae (roughly four da ys), or parasitoid (Figure 2-5). Data were then collected on whether butterfly larvae or a parasite emerged. Larval Predation The final part of the experim ent was to te st predation effects on larvae of Monarch butterflies ( Danaus plexippus ) in north central Florida and Common Morpho butterflies ( Morpho peleides ) in Costa Rica. For each location, two host pl ant rows were placed radiating out from

PAGE 37

37 the farm epicenter, much like the first experime ntal arrangement descri bed earlier. However, since there were only two host plants per row, one was placed at the edge of the epicenter perimeter (distance = 0 meters) and the othe r was placed 100 meters away (distance = 100 meters) (Figure 2-6). For each plant at the north centr al Florida location, five 3rd-instar Monarch butterfly larvae were directly placed on the plant a nd allowed to feed in a controlled laboratory environment for 24 hours (Figure 2-7). This st ep was to acclimate the larvae on the host plant and to make sure that they were not prone to wa nder off of the plant within the 24 hours that the trial was actually running. After this 24 hour acclimation period, the plants were taken from the laboratory and placed in their appropriate positio ns in the field. After another 24 hours, data were collected on the number of larvae remain ing for each plant. This entire process was replicated twice for each nor th central Florida location. For the Costa Rica location, two rows of host plants of the Common Morpho ( Morpho peleides) were placed in an identical fashion as th e plants at the north central Florida location (Figure 2-6). Because the host plants were trees and relatively large, they would not be able to fit into a controlled laboratory room, as the Monarch larvae were. Th erefore, an alternate method was used. With the potted host plants already in the field, five 3rd instar Common Morpho larvae were placed directly on the plant and then encl osed in a predator screen (Figure 2-8). This screen was a fine mesh bag tied at both ends (using a secure overhand reef knot), both preventing the larvae from wandering as well as excluding parasites and pr edators from attacking the larvae while they were bei ng acclimated to the host plant (F igure 2-9). These screens are used in the very same way by the butterfly farm and have provided successful results throughout

PAGE 38

38 the years, according to the farm manager. After 24 hours, the predator screens were removed. Then 24 hours later, data were collected on th e number of larvae remaining on each plant. In addition, a control variable was used for th is experiment at all geographic locations. At a separate location away from the Butterfly farm, in similar habitat, a false epicenter was marked with marking flags. This epicenter was the same size and dimensions as its experimental counterpart, but did not have any butterfly farming operations taking place within its borders (Figure 2-10). Two rows of host plants were set up in an identi cal fashion as described for the experimental variable. The researcher oriented the rows so that both experimental and control variables were along the same co mpass direction. Acclimation a nd data collection procedures were identical to the procedure us ed for the experimental variable.

PAGE 39

39 Figure 2-1. Entrances to the th ree butterfly farms used throug hout the experiment. Clockwise from upper left, entrances to El Bosque Nuevo, Shady Oaks, and Greathouse butterfly farms Table 2-1. List of butterf ly and host plants used Butterfly Species for Costa Rica Host plant Species for Costa Rica Heliconius sapho Passiflora auriculata Caligo memnon Heliconia sp. Morpho peleides Leguminosa cristobal Butterfly Species for Florida Host plant Species for Florida Danaus plexippus Asclepias curassivica Junonia coena Plantago lanceolata Papilio polyxenes Petroselinum crispurn

PAGE 40

40 Figure 2-2. Diagram of farm epicenter (green ci rcle) with radiation rows of hostplants at 10 meter transects (yellow dots). Transects re peated for each row, but shown here for only one row. Three different host plants were used, corresponding with the three differently colored rows in this diagram

PAGE 41

41 Figure 2-3. Dense jungle causing one row of host plants to be shifted slightly

PAGE 42

42 Figure 2-4. Several plants needed to be placed close to roads in the Florida experiments

PAGE 43

43 Figure 2-5. Screened vials and st ereoscope used to observe the em ergence of parasites or larvae Figure 2-6. Diagram of larv al predation experiment

PAGE 44

44 Figure 2-7. Controlled environment used to acclimate Monarch larvae to their host plant

PAGE 45

45 Figure 2-8. Morpho peleides host plant Leguminosa cristobal before and after predator screen application Figure 2-9. Close-up view of predator screen

PAGE 46

46 Figure 2-10. Control plot used in north central Florida experiments

PAGE 47

47 CHAPTER 3 RESULTS Main Experiment Butterfly Oviposition in North Central Florida After a period of 10 days of data collection at tw o north central Fl orida butterfly farms, 60 observations (30 for each farm) were taken on oviposition for each of 11 transects. At these two farms, sufficient data were collected from Monarch butterfly oviposition, but neither farm yielded sufficient oviposition data for the Eastern Black Swallowtail or the Common Buckeye. Although Buckeyes and Swallowtails were freque ntly seen at both farm s by the researcher, no Buckeye eggs and only 2 Swallowtail eggs were observed during the en tire 10-day period on all transects. Therefore, statistical models and summary statistics are meaningless for these two butterflies. Monarch oviposition, however, was indeed commonplace, and therefore is the main subject for this oviposition experi ment in north central Florida. Mean egg number per day was highest for thos e transects closest to the farm epicenter. Then, egg number proceeded to decl ine and level out at a lower mean for transects more distal to the farm epicenter (Figure 3-1). Figure 3-1 prov ides a summary of mean egg numbers per day at each transect for monarch oviposition at the north central Florida farms. Summary statistics are provided in Table 3-1. From th ese data, it is evident that the highest egg numbers are found closer to the farm, with transects 1, 2, 3, and 4 yielding the highest egg numbers (3.033, 2.933, 2.933, and 2.700 eggs/day respectively). Starting with transect 5, which exhibited an average of 1.683 eggs/day, egg numbers then dropped off by approximately 1.5 eggs per day. Transects 5 through 7 exhibit a downward sloping trend as egg numbers proceeded to decline until rising again at transect 8. However, transect 9 exhibited the lowest egg number of all, at 1.067 eggs/day. Then, a slight increase was exhibited with transect 10 seeing 1.4 eggs/day and transect 11 seeing 1.633 eggs/day. This up and down pattern between 5 and 11 is only slight when

PAGE 48

48 compared to the drop from 1 to 5. From these da ta, there appears to be a noticeable difference in egg numbers between transects in regards to distance from the farm epicenter. Looking at the data summarized in Table 3-1, we see that the highest standard deviations correspond with the highest mean egg numbers. These high standard deviations wreak havoc on statistical tests, lessening any significance that the data may inhe rently possess. Consequently, we cannot say that there is a significant diffe rence in egg averages between transect 1 and transect 11, much less between other more closelyspaced transects. However, when we look at the 3rd quartile data (Table 3-1), ther e is a clear increase in freque ncy of higher egg numbers with transects more proximal to the epicenter, sp ecifically transects 1 through 4. Additionally, median values are significantly higher with tran sects 1 through 3. Perh aps due to the sample size, we cannot prove significant differences between any of the transects using summary statistics alone. The next section will detail more complicated statistical analyses and findings. In order to further analyze butterfly ovipos ition, the researcher used the model of repeated measures over time, treating the experi ment as a randomized block design with day, location and transect as the fact ors. This allows the research er to put the fixed effects of location, day and transect into blocks in order to measure how influential each fixed effect is to the variance in egg number. The raw data were then transformed (i.e ., egg numbers) using a log+1 transformation in SAS to standardize the residuals in order to meet the assumptions of the model. The transformed data were then analyzed in SAS using a proc mixed procedure which generated the significance st atistics found in Table 3-2. Variations in egg number can be attributed to 4 different effects: location, transect, day, and the interaction between location and day. There were no reported interaction effects between location and transect, nor day and transect Table 3-2 shows a list of the effects and

PAGE 49

49 their corresponding F-values and P-values, all at pres et al pha = 0.100. Transect effect was found to have a p-value of 0.2049, which is slightly larger than the pre-s et al pha level. However, this is still a relatively low value. In conjunction wi th the summary statistics described previously, there is enough evidence to susp ect that the transect number (i .e., distance from the epicenter) has an effect on butterfly oviposition. Due to the strong effect that location and day had on egg number, it merited further examination of the data based on individual loca tions and individual days See Figure 3-2 and Figure 3-3 for separate graphs for average egg nu mber per day at each location in north central Florida. From the data shown in Figure 3-2 and Figure 3-3, we can s ee a definite difference between locations. Shady Oaks has a more pronounced negative co rrelation between egg number and transect, while for Greathouse, egg number drops sharply ar ound transect 5 and 6, but then rises from 6 to 11 (Figure 3-2 and Figure 3-3). As mentioned before, there is an interacti on effect between location and day. Figure 3-4 shows the egg number broken down by individual days as well as the individual farm. This interaction means that the combination of day a nd location yields more variance in egg number than simply the sum of their individual main ef fects. We see that Shady Oaks yielded overall fewer numbers of eggs in a vacillating fashion, while Greathouse exhibited a bell-shaped pattern, with most oviposition occurring on day 5 and 6, with a sharp drop off on day 7 through 10. The day effect was quite pronounced as well. Most eggs were deposited during the middle of the experiment, with days 4, 5 and 6 seeing the most eggs at 201, 189 and 252 eggs respectively (Figure 3-5). Then, on day 7, a sharp decline in oviposition was observed, which then continued until the end of the experiment.

PAGE 50

50 Finally, the location effect is shown below in Figure 3-6. There was over a 3-fold increase in total oviposition fr om Shady Oaks to Greathouse, with 288 and 997 eggs deposited, respectively, over the 10-day period. Main Experiment: Butterfly Oviposition in Costa Rica Oviposition in the Costa Rica trial was very surprising, as absolutely no eggs were oviposited on any of the hostplants during the entire 10-day duration of the experiment. Although this was an unexpected result, it does ra ise some interesting questions, which will be addressed in the Discussion chapter. Larval Experiment: Larval Survivorship as Dete rmined by Distance from Farm Epicenters in North Central Flor ida The larval predation experiment suggests that there is indeed an effect that the butterfly farm has on its environment, but perhaps not in the same way as was originally predicted. After running the experiment at the farm and running a control in a field of similar habitat approximately 5 miles away from the farm, severa l comparisons were made. First, total larval survivorship (i.e., not accounting for distance from the epicenter) at the experimental plot was compared to a control plot. Higher levels of su rvivorship among larvae at the experimental plot were found when compared to the control plots (Figure 3-7). Surv ivorship at the experimental plot was 41%, while survivorship at the control plot was 14% (T able 3-3). Using a T-test at alpha = 0.005, the researcher was ab le to conclude that there is a significant difference between the experimental and control plots, in terms of larval survivorship. Second, it was determined that there was also a significantly higher level of survivorship for larvae situated closer (i.e., located at transect station 1, which was adjacent to the perimeter of the farm epicenter) to the farm epicenter, when compared to larvae situated farther (i.e., located at transect station 11, which was 100 meters away from the perimeter of the farm epicenter) fr om the farm epicenter

PAGE 51

51 (T-test at alpha = 0.005). An averag e of 58% larval survivorship on plants in the near location is significantly greater than the 25% larval survivorship at the far location (Figure 3-8). Lastly, larval survivorship between near and far locations in the control plot was found not to differ significantly (Table 3-5). Larval Experiment: Larval Survivorship as Dete rmined by Distance from Farm Epicenters in Costa Rica Identical to the north cen tral Florida farm s experiment, a control plot was run in conjunction with the experimental farm plot. Af ter collecting the data, several comparisons were made. First, using a T-test at alpha = 0.01 (lar ger alpha due to smaller sample size), it was determined that there was no significant differe nce in total survivorship between control and experimental plots (i.e., not accoun ting for distance from the epicenter). Percent survivorship for the experimental plot was 83% versus 63% fo r the control plot (Table 3-3). The second comparison was between near and far locations for larval survivorship in th e experimental plot. With near and far points exhibiting 90% and 75% larval survivorship, re spectively, a T-test at alpha = 0.01 concluded that ther e was no significant difference be tween the two di stances (Table 3-4). However, a T-test at alpha = 0.01 did find a significant difference in larval survivorship when comparing near and far treatments in the cont rol plot (Figure 3-9). Percent survivorship in this control plot was found to be 40% for the near plot and 85% for the far plot (Table 3-5). Egg Parasitism The egg parasitism study in north central Fl orida generated significant data. However, the findings were very unexpecte d. By collecting a total of 12 e ggs per transect, without regard to day, none of the collected eggs yielded any evidence of parasitism. One hundred percent of eggs collected from all transects hatched to produce live Monarch larv ae. Due to the limited

PAGE 52

52 number of eggs deposited at transects 9, 10, and 11 at Shady Oaks, the full 12 replications were not able to be conducted. Rather, the sample size was 6, 9, and 6, respectively.

PAGE 53

53 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Transect (at 10m intervals) Egg Number 3.0332.9332.9332.7001.6831.2831.1831.5671.0671.4001.633 1234567891011 Figure 3-1. Summary graph of aver age egg number for each transect Table 3-1. Summary statistics for Nort h central Florida oviposition experiment Transect Sample size Mean egg number St Dev Qmax Q3 Median Q1 Min 1 60 3.033 3.835 15 6 1 0 0 2 60 2.933 3.901 14 4.5 1 0 0 3 60 2.933 4.137 14 4 1 0 0 4 60 2.700 4.777 21 3 0 0 0 5 60 1.683 2.861 13 2 0 0 0 6 60 1.283 2.387 10 1 0 0 0 7 60 1.183 1.987 9 2 0 0 0 8 60 1.567 3.061 12 2 0 0 0 9 60 1.067 1.812 7 2 0 0 0 10 60 1.400 2.688 11 2 0 0 0 11 60 1.633 3.650 17 1.5 0 0 0 Table 3-2. Significance values obtained from the SAS proc mixed procedure ( =0.1) Effect F-Value P-Value Location 32.18 <.0001 Transect 1.42 0.2049 Day 13.64 <.0001 Location*Day 24.25 <.0001

PAGE 54

54 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0Transect (at 10m intervals) Egg Number 4.6004.1004.2334.3002.6331.5331.8672.4671.9332.5003.067 1234567891011 Figure 3-2. Mean egg number at each transect for Greathouse Butterfly Farm 0.0 0.5 1.0 1.5 2.0 Transect (at 10m intervals) Egg Number 1.4671.7671.6331.1000.7331.0330.5000.6670.2000.3000.200 1234567891011 Figure 3-3. Mean egg number at each transect for Shady Oaks Butterfly Farm

PAGE 55

55 0 20 40 60 80 100 120 140 160 180 200 Day Shady Oaks 60902806319285526 Greathouse 66101158173189189853033 12345678910 Figure 3-4. Location and day affected egg number more than just the su m of their individual roles 0 50 100 150 200 250 300 Day Total Egg Count 126110158201189252104585829 12345678910 Figure 3-5. The day effect on egg number was rela tively bell-shaped, with a peak in egg counts for days 4 through 6 after initial placement of the potted plants

PAGE 56

56 0 200 400 600 800 1000 1200 Location Egg Number 288 997 Shady Oaks Greathouse Figure 3-6. Location had a pronounced effect on e gg number, with almost 3 times as many eggs at Greathouse when compared to Shady Oaks Table 3-3. Mean % survivorship among larvae at experimental and control plots for all geographic locations Farm Experimental Control Shady Oaks 43 18 Greathouse 40 10 Florida combined 41 14 El Bosque 83 63 Table 3-4. Mean % survivorship among larvae at far and near distances in the experimental plot for all geographic locations Farm Far Near Shady Oaks 25 60 Greathouse 25 55 Florida combined 25 58 El Bosque 75 90 Table 3-5. Mean % survivorship among larvae at far and near dist ances in the control plot for all geographic locations Farm Far Near Shady Oaks 35 0 Greathouse 0 20 Florida combined 18 10 El Bosque 85 40

PAGE 57

57 0 10 20 30 40 50 Treatment % Survivorship 14 41 Control TreatmentExperimental Treatment Figure 3-7. Mean % survivorship among larvae at experimental and co ntrol plots for north central Florida farms 0 10 20 30 40 50 60 70 Distance % Survivorship 58 25 Near Far Figure 3-8. Mean % survivorship among larvae at far and near dist ances in the experimental plot for north central Florida farms

PAGE 58

58 0 20 40 60 80 100 Distance % Survivorship 4085 near far Figure 3-9. Mean % survivorship among larvae at far and near dist ances in the control plot for the northern Costa Rica farm

PAGE 59

59 CHAPTER 5 DISCUSSION Throughout the world today, butterfly farm ing is becoming increasingly common as a source of revenue or commerce for communities in both developed and de veloping nations. It has been reported that this indus try can generate six-digit annual revenues in even the most rural of communities in developing nations (Gordon & Ayiemba 2003). While specializing in butterfly rearing and breeding, many of such farm s also provide educatio nal opportunities for the surrounding communities, which serve to educate the public on farming practices as well as butterfly and habitat co nservation. By providing economic incentives to the local community and actively engaging in nature education a nd conservation, butterfly farming is rapidly becoming a part of ecotourism and c onservation efforts across the globe. With the emergence of butterfly farming operations, so too emerges a need to better understand factors affecting butterfly natural history around the farm. It was the objective of this study to better understand the impacts of farms on butterfly abundances, using oviposition as an indicator, butterfly larvae predation rates, a nd butterfly egg parasitism, all with regard to proximity to butterfly farms. Butterfly Abundance and Oviposition at North Central Florida Farms Monarch butterfly oviposition was, on average, in greater numbers at transect stations 1 through 4 when compared to transect stations 5 through 11. Although there is not a smooth downward sloping trend in egg number as one goes from transect stations 1 to 11, we can say that Monarch butterflies are much more likely to be found ovipositing on plants within 40 meters of the butterfly farm epicenter. When examining transect stations 1 through 4, specifically, we observed a completely negative correlation between egg number and transect station number as one goes from transect station 1 to trans ect station 4, which was hypothesized to occur

PAGE 60

60 throughout the entire range of transect stations 1 through 11. These data ultimately suggest that butterflies are more abundant in th e areas directly adjacent to the butterfly farm epicenter, and are correspondingly less abundant farther from the farm epicenter. This idea will be discussed more later, but for now, it is necessary to stress the fact that there was a marked difference between transect stations 1 through 4, and transect stations 5 through 11. The difference between 1 through 4 and 5 thr ough 11 is likely due to the researchers placement of the farm epicenter perimeter. This imaginary perimeter line was placed in order to designate the perimeter of the actua l working part of the farm. It was designed to incorporate all the hypothesized attractive forces, such as host pl ants, nectar plants, re aring cages, butterfly gardens, and laboratories that could potentially influence oviposition pr eference by containing resources used by female butterflies. Because there were multiple areas with host and nectar plants throughout each of the farm epicenters, some of the less attractive and less resourceloaded components of the farm epicenters, such as closed laboratories and more shielded plant nurseries, may have acted as a buffer between the perimeter and the more attractive rearing cages and host and nectar plant nurseries. If wild populations of butterflies are being attracted principally by these interiorly located plants and/or the associated butterflies, they may fly directly to these areas and subsequently limit future flight s to a set radius around these host and nectar plants, while searching for alternatives depending on their oviposition preference. Perhaps the 30 meter transect station (at station #4) approached the outer limit to their searching behavior, rendering transects 5 through 11 less obvious. According to Renwick & Chew (1994), learning plays a vital role in the recognition of suitable sites for oviposi tion. Thus, it could be possible for butterflies to have learned the location of the best host plants within the farm epicenter prior to the start of this experiment having already assigned them oviposition priority.

PAGE 61

61 Previous research suggests that when female butte rflies are confronted w ith an array of potential hosts, they exhibit a hierarchy in preferences, laying most eggs on the preferred plant, and fewer on the next, and so on (Courtney et al 1989, Thompson 1988). A second scenario possibly explaining the diffe rence between station groups 1-4 and 5-11 could be that egg number data (i.e ., oviposition) may have come fr om butterflies from within the farm rather than wild butterfli es. These within farm butterf lies could be ones that have escaped from the enclosures, or possibly hatched from eggs that remained undetected by farm workers as the host plants are moved around the farm. Oviposition from escapee and other farmraised butterflies may be more contained to the interior of th e farm epicenter, depending on location of the particular reari ng cage from which the butterfly escaped or the host plant from which the butterfly emerged. Previous studies have found that visual conspicuousness of host plants is related to increased oviposition (Courtney 1982). Therefor e, plants located at closer transects may be more conspicuous to the esca pee and farm-raised butte rflies, and would likely exhibit higher levels of ovipos ition. Although this may apply to the steady decline in egg number from transect stations 1 through 7, it does little to explain the drastic drop from transect station 4 to transect station 5. The conspecific attraction hypothesis states that high-density popul ations yield higher immigration rates to, and lower emigration rate s away from, the habitat in question (Ray et al 1991). Should these escapee butterflies be conten t with the available re sources of the farm, many may reside inside the farm epicenter. Th e likelihood of wild butte rflies being present in the outdoor nurseries and butterfly gardens, in addition to these escapee butterflies, could serve as motivation for butterflies to remain in th e area, according to the conspecific attraction hypothesis.

PAGE 62

62 In this set of experiments examining butterfly abundance an d ovipostion around north central Florida butterfly farms, there were thr ee main effects: day, location, and transect. The following sections will discuss possible expl anations for these results in detail. Main Effect 1: Day The effect that the number of days host plant exposure to gravid females in the vicinity had on egg number throughout the experiment was ve ry significant (p-value < .0001). With days four, five, and six exhibiting the most oviposition, it is plain to see that butterfly abundances were highest during the middle of the experiment, forming a bell curve. The most probable explanation for this initial spike is that the fi rst few days were required for the butterflies to locate, accept, and learn the location of the host plants, which resulted in high oviposition levels on the subsequent days. Then, once the host plants became heavily used and environmental conditions stressed the plants, th ey became less suitable to gravid females, resulting in less frequent oviposition. Butterflies rely on specific sensory cues in searching for appropriate host plants, with visual perception playing a major role (Prokopy & Owens 1983, Singer 1993). Using visual perception, butterflies overwhelmingly use the shape and color of the host plant as a search tool, with the color and shape of leaves particularly important (Stanton 1984). According to Courtney and Forsberg (1988), landing freque ncies by gravid female butterflies have been seen to depend on the relative abundance of the ho st plant in the area. With a wealth of competing host plants in the genera l vicinity, all providing similar vi sual cues, it is possible that some gravid females did not detect the expe rimental host plants immediately. As these experimental host plants were detected by more and more gravid females, the oviposition rate increased until reaching a peak -in the middle of the experiment. Once the plants elicited the maximum oviposition rate during the temporal median of the experiment, oviposition began to dec line sharply. It is possible th at during the days that these

PAGE 63

63 plants were in the field, they became degraded due to high levels of oviposition and stressful environmental conditions such as intense sunlig ht, high temperatures, ha rd rainfall or wind dehydration of leaves. A study by Ilse (1937) was th e first to report that females in most families of Lepidoptera engage in a drumming action on the surface of the leaf on which they intend to oviposit. Later described as a scrapping actio n by Fox (1966), Fox suggested that the spines on the foretarsi abraded the leaf surface in order to release essential oils from the leaf, which were likely used in host plant selection and accept ance. However, this supposed degradation of the leaf surface is not limited to just female lepidopterans. Some male danaid butterflies have been reported to scratch withered leaves in order to attain pyrr olizidine alkaloids from the plant (Boppre 1983). Two studies, one by Schurr & Holdaway (1970) and the other by Renwick and Radke (1982), suggest that injure d and stressed plants may actua lly stimulate the release of volatiles from host plants, which then act as oviposition deterrents to several species of Lepidoptera. Should butterflies be instigating their hos t plant to release an oviposition deterrent caused by plant damage and stress, we would exp ect high levels of oviposition to be followed by a sharp decline, as seen in this experiment. Environmental conditions may have also ta ken their toll on the pl ants, possibly rendering them relatively unsuitable for oviposition. This experiment was conducted in August, which is typically a very hot month in Florida, and it is not uncommon to see temperatures above 100 Fahrenheit. Temperature readings during this experiment were consis tent with historical records, and temperatures in the high 90 s were frequently observed. Su mmer storms are also common, bringing strong wind gusts as well as flooding rain. Although thes e plants were cared for daily, many were in direct sunlight for the majority of daylight hours and also exposed to the daily pounding rain. It was evident th at as time went on, some plan ts were showing signs of

PAGE 64

64 degradation. As stated previously, color of the host plant plays a major role in whether or not a female butterfly is successful in finding an acceptable host (Prokopy & Owens 1983). As stated by Myers (1985) and Wolfson (1980), color may de pend on plant chemistry, which is directly related to the physiological condition of the plan t. It is possible that as the experiment progressed, the physiological condition of the pl ants gradually rendered them less and less visible, attractive, and suitable to gravid female butterflies. Conspecific egg loads have been found to be oviposition deterrents for many species of butterflies, including those in the genus Danaus (Rausher 1979b, Rothschild & Schoonhoven 1977). For this very reason, eggs were removed daily in an effort to minimize oviposition deterrence. However, some species of Lepidopt era have been known to chemically detect the presence of conspecifics (Renwick & Radke 1980) It is possible that residual chemicals from the egg remained on the leaf surface, which ma y have acted as an oviposition deterrent. Despite these possible explanat ions for a bell-shaped curve in oviposition for day and the very significant p-value, we must not overlook the fact that this bell-shaped curve was observed solely at Greathouse. Because Greathouse yielded significantly mo re data than Shady Oaks, the analysis of the combined data overemphasized the results from Greathouse. However, there was clearly a pronounced day effect at Shady Oaks as well, with fluctuations up and down throughout the 10-day trial. Analysis of the day main effect suggests that more trials need to be conducted if we are to confidently say that a be ll-shaped pattern is indeed occuring. Main Effect 2: Geographic Location The effect that geographic location had on egg number was also very significant (p-value < 0.0001). When the data are analyzed by locati on, we see that the Greathouse trial yielded approximately three times more eggs overall than the Shady Oaks trial. Similarly, almost three

PAGE 65

65 times more eggs were found, on average, for each transect. For certain days, almost 10 times more eggs were found at Gr eathouse versus Shady Oaks. One possible reason for this difference is th e farm epicenter size, which was observed by the researcher to be directly pr oportional to the number of nurser ies, butterfly rearing houses, and other parts of the farm involved in butterfly production. The farm epicenter for Shady Oaks was nearly 7,000 square meters, while that of Great house was just over 12,600 square meters. With twice the total area, occupied by nectar plants host plants, and butte rfly enclosures at Greathouse, we see that there may be a positive co rrelation between the size of the butterfly farm and rates of butterfly abundances around the farm The greater size and number of butterfly enclosures could increase the potential for but terflies escaping on the principle that more enclosures create more risk of escape. A second difference between the two farms was the landscape su rrounding the farm epicenter, and there were two major differences. First is the amount of forested land surrounding the farm. The type of forested land was similar at both farms, consisting of mostly pine and oak trees, with some undergrowth, all growing in sandy, well-drained soil. Shady Oaks was bordered by far more forested areas with all but two ro ws traversing part of the surrounding forest. Although the majority of host plants were placed in the grassy ar eas that occupied most of the experimental plot, some plants were placed in th e forested land, rendering them much less visible to gravid female butterflies (Figure 4-1). Conversely, Greathouse butterfly farm was situated in a way that only one row of hostplants traversed fore sted land. The vast majority of plants were situated in well-kept grass, allowing the visu al cues to be better perceived by searching butterflies (Figure 4-2). The s econd landscape characteristic th at may have led to greater oviposition rates at Greathouse was the presence of a residence within close proximity to the

PAGE 66

66 butterfly farm epicenter (Figure 4-3). At this residence was a butterfly garden. Filled with abundant nectar sources and cove ring an area of approximately 200 square meters, this garden was a butterfly attractant, as observed by the re searcher. Although only one row of host plants traversed this area, high rates of oviposition was observed for t hose transects proximal to the garden in this particular row. Although this re sidence and garden were very close to the farm epicenter, they could not be counted as part of the epicenter because it was not involved in the operations at the butterfly farm. One more thing to note about the Greathous e trial was the presence of a large monarch butterfly enclosure, filled with hundreds of m ilkweed plants and adult and larval Monarch butterflies located at the very edge of the farm epicenter (Figure 44). It just so happened that one of the Monarch butterfly hos t plant rows used in this ex periment was placed directly adjacent to this enclosure, as dictated by the ra ndomization scheme designed prior to the start of the experiment. Not coincidentally, the first th ree host plants radiating out from the epicenter here also exhibited the highest number of Mona rch eggs throughout the entire experiment, with some days seeing upwards of 20 eggs on each plan t. High oviposition levels here are possibly the result of high plant visibility, and may have been attr acting gravid females. When the females arrive to oviposit on what they feel will be suitable habitat, they are excluded from the enclosure and choose to oviposit elsewhere. With several of the e xperimental host plants located in close range, they become a probable choice fo r the gravid females. Although these plants did have relatively high levels of oviposition, these transect stations were not outliers in terms of overly high levels of oviposition when compared to the other data from this study. Oviposition rates were affected by the number of days the plants at these tr ansects were exposed to gravid

PAGE 67

67 females in the vicinity just as other plants were that were not lo cated near this Monarch butterfly enclosure. Main Effect 3: Transect Station The effect of transect station on oviposition wa s not as statistically significant as either day or location, but still receive d a relatively low p-value (p-v alue = 0.2049) and interesting summary statistics. From the data, we found that there is a positive correlation between high oviposition levels and proximity to butterfly farm epicenters. With ovipositi on also serving as an indicator for butterfly abundance, the question must be posed as to why butterflies would tend to be more abundant within 40, perhaps 30, meters from the perimeter of the farm epicenter? As stated by Rausher (1979a), females that oviposit on plants yielding low larval survivorship or poor larval grow th will likely leave fewer desce ndents than females that oviposit on the more appropriate host plants. For this r eason, the choice that a female butterfly makes on where to oviposit is critical. Although exception s have been found, it is widely accepted that oviposition search behavior has evol ved to ensure that females lay eggs in the habitats that are most suitable for the growth and development of the juvenile stag es (Gilbert & Singer 1975, Wiklund 1977). Therefore, gravid females may prefer the plants closer to the farm epicenter because these plants are perceived by the female to be more suitable for development of their offspring. Evidently, as the butterflies eval uated the balance between positive and negative signals from the plant and external stimuli, such as presence of eggs and other insects, the plants closer to the farm epicenter are more suitable. Although unproven here, it is possible that the farm may be acting as a magnet, both attracti ng butterflies from wild populations, as well as retaining any farm-raised or escapee butterflies from within the farm itself.

PAGE 68

68 Each of the farms tested in this experime nt report that they have thousands of adult butterflies present at any one time, as well as t housands of host and nectar plants within the borders of the farm epicenter. With high levels of butterflies present for breeding at the farm, as well as the probability of host and nectar plants se rving to attract even more butterflies, we can say that there are high conspecific densities among butterflies being reared at the farm. In turn, these high conspecific densities could themselves serve as attractants, according to the conspecific attraction hypothesis, further increasing levels of conspecific densities (Ray et al ., 1991). According to Odendaal et al (1989), butterflies are commonly attracted to areas of high conspecific densities due to the increased probab ility of finding a suitable mate. Additionally, Gilbert and Singer (1973) suggested that a high density of butterf lies in an area, es pecially of the same species, could be used by a butterfly as an indicator for suitable habi tat. The conditions at the farm in which high densitie s of butterflies are continually present is a possible force of attraction that may cause high levels of ovipositi on closer to the farm epicenter. With the epicenter providing host plants, n ectar plants, and possibly mating opportunities for butterflies, it is not surprising that there was less oviposition on those plants further away from the farm. This study, however, did not exam ine butterfly abundance within the farm epicenter. There is a need for further research to be done regarding the populations of butterflie s actually inside the butterfly farm epicenter. Although the results from this study indicate that the farm may serve as an attractive force, the data can only suggest that butterflies are likely to be more abundant closer to the farm epicenter. Further studies are need ed to ascertain whether or not higher butterfly abundance is caused by attraction, and if so, to what degree. Larval Predation in Florida and Costa Rica A very unexpected result of this study was th at the presence of the butterfly farm, in all geographic locations tested, provided for higher levels of larval survivorshi p. This is true when

PAGE 69

69 comparing the experimental plot with the contro l plot, as well as when comparing the near variable with the far variable at the experimental plot. Higher levels of predation on larvae that are situated further from the butterfly farm could be the result of either a higher presence of pred ators or higher rates of larval detection by predators. It could also be possible that th ere are multiple populations of predators surrounding the farm, with separate populations predating upon the larvae at different di stances from the farm epicenter. Perhaps the predator populations situat ed closer to the farm predate mainly on larvae in and around the farms gardens, nurseries, etc ., and were satiated wit hout needing to predate upon the larvae presented to them in this experi ment. In addition, it is possible the predator population away from the farm, not typically ha ving easy prey such as the larvae that were placed in this experiment, took advantage of the opportunity to predate upon these easy prey items. It could have also been possible that different species of predators were at the different distances from the farm. Perhaps high predation rates at the contro l plot were due to one species, while predation at the farm plot was from another species. The control plot in Florida, although similar habitat to the farm plot, was unmowed, brushy and in very sandy soil. Solenopsis ants were observed here and could have been the main predator involved. Similarly, predation at the farm could have differed depending on the spec ies of potential predator as well. Although Solenopsis ants were also found here, they could ha ve been more contained due to the mowing regime (approximately once a week), rendering them a subordinate predator. Lastly, relatively high levels of human activity could have aff ected predation levels. It is possible that the near constant presence of eco tourists and farm workers around the farm grounds was enough to impose a threat to avian and lizard predators (Figure 4-5) By trying to avoid humans, potential predators would be indirectly deterred away from the farm. Should this be

PAGE 70

70 true, it is a unique example of the benefits of integrating ecotourism (e specially the high energy and noisy student groups) into a butterfly farming operation. Although statistical T-te sts indicate that Costa Rica did not follow the same pattern of significant differences for north central Florida, the raw data clearly follow the same pattern as north central Florida. For this reason, the researcher does not feel that there is any notable difference between the Costa Rica and north central Fl orida trials. In order to test for significant differences between the localities, additional studies are needed. There are numerous possibilities for why an overall decline in pred ation was observed as one gets closer to the farm epicenter, but we cannot ignore the plain fact that there were statistically significant differences allowing us to make this conclusion in the first place. There is a great need for further rese arch to be done on predation rate s in regards to proximity to butterfly farms. In particular, future studies should concentrate on causes for these unexpected results. Egg Parasitism in Florida and Costa Rica Egg parasitism was found to be non-existe nt among monarch butterf ly eggs during the 10-day experiment at all three geographic locatio ns in Florida and Costa Rica. Although this bodes well for the prognosis of ove rall parasitism rates ar ound these particular butterfly farms, we cannot make any general conclusions from th is experiment. These results certainly do not mean that parasites and parasitoids are absent at butterfly farms, and they likely do not mean that they are necessarily absent from the fa rms used in this experiment. What is likely is that the experiment was not designed in a way to attract pa rasites and parasitoids to the butterfly eggs. The first problem is that the butterfly eggs may not have been available to the parasitoids for long enough. In a study by Fatouros et al (2005), wasps were found to be more attracted to leaves with eggs deposited for 48 hr, rather than eggs for 24 hr. Fatouros et al (2005) also found

PAGE 71

71 that eggs deposited for 72 hr elicited and even more significantly stronger response and suggested that substances associated with egg deposition could have diffused into the leaf tissue or the leafs wax layer, triggeri ng a chemical response by the leaf. This chemical, in turn, would at very least arrest, and possibly attract parasitoids. In my ex periment, eggs were left on the plant for no more than 24 hr, which may not have been enough time for synergistic effects between the plant and the egg to attract parasitoids. The reason that eggs were removed daily was to prevent oviposition competition in the oviposition study. Research has suggested that gravid females may pass up suitable host plants du e to the presence of conspecific eggs (Rausher 1979b, Rothschild 1977). The second problem is that the plants were spaced, from a parasi toids perspective, relatively far apart from each other. This pr evented parasitoids from dispersing by means of walking and jumping, which is a vital means of locomotion for butterfly parasitoids (Pak et al 1985). While nurseries inside the farm epicenter provided for high connectivity between host plants, the 10-meter spacing used in the experiment provided re latively no connectivity. Gingras and Boivin (2002) stated that connectivity betwee n host plants and host plan t parts best explains variability in parasitism rates. Due to these reasons, it is likely that parasito id presence is concentrated to areas within the farm epicenter. Areas in which eggs are left on host plants for greater than 48 hrs, as well as high levels of plant connectivity and homogeneity, are likely to present ideal habitats for populations of egg parasitoids, thus there was little need for them to disperse to the outer areas of the farm (Lukianchus & Smith 1997). It would be beneficial to conduct further studies from within the farm epicenter and select specific habita ts and specific groups of plants rather than the entire farm, as was done in this experiment.

PAGE 72

72 Butterfly Abundance and Oviposition at the Northern Costa Rica Farm The results from the butterfly abundance st udy in Costa Rica were unexpected and very surprising. With the number of eggs deposited in the north central Flor ida trials ranging from several hundred to almost one thousand during th e 10-day experiment, finding absolutely zero eggs during the Costa Rica trial is perhaps as interesting as if there were h undreds of eggs found. There are several factors that may contribu te to this overall l ack of oviposition. One explanation has to do with the extreme competition among host plan ts in the tropical environment. As stated before, butterflies make tremendous use of visual stimuli to locate host plants, particularly the color and shape of the host plants leaves (Stant on 1984). Furthermore, relative abundance of host plants in an area ha s been shown to affect landing frequencies by gravid female butterflies (Courtney & Forsberg 1988). Consequently, the tremendous diversity of plants in the rainforest, some as host plants for these species, others with similarly shaped and colored leaves, could have acted as competition towards the potted plants placed out in this experiment. With the experiment lasting 10 days perhaps this was too short of a duration for butterflies to recognize and learn the location of su ch host plants in a sea of both nearly identical and intricately complex stimuli. Seasonality may have played a part in the vi sual homogeneity of th e landscape as well. This experiment was conducted during the rainy s eason in Costa Rica and, thus, there was little to no flowering exhibited by the plants. Acco rding to Janzen (1967), tropical plants in Guanacaste, Costa Rica, flower and fruit in the dr y season due to the fact that investing into nonvegetative activity is least detrimental to their vegetative competitive ability when rains are less frequent. The presence of flowers may have in creased oviposition by prov iding another visual cue for the butterflies, which may have yielded oviposition on the host plant.

PAGE 73

73 The host plants used in this experiment were the same species used by the farm for breeding their own butterfly stock. However, these host plants could have been the wrong choice, nonetheless. Landing frequencies by gravid female butterflies, and ultimately oviposition rates, have been seen to fluctuate with the seasonal changes in host plant abundances (Rausher 1979a). It is possi ble that while farm-raised butterflies will oviposit on these host plants year-round, wild bu tterflies may be seasonally adapted to searching for different species during the time this experiment was conducted. This could also apply to the north central Florida experiment, providing ex planation for why there was l ittle to no oviposition by both Eastern Black Swallowtails and Common Buckeyes. The layout of the Costa Rica farm posed a particular problem fo r measuring oviposition of Caligo memnon. Near the farm epicenter, but outside of the operational part of the farm, there was a huge stand of Heliconia plants. These were the same species used at the farm to rear their own Caligos and consequently, the same species used in this experiment (Figure 4-6). With hundreds of Heliconia plants very near two of the experimental Heliconia rows, the hostplant competition here for visiting female Caligo was especially high. Throughout the beginning phases of the expe riment, there was a very high abundance of Anartia fatima butterflies around the farm. As the expe riment progressed, and day after day no eggs were found at any of the host plant transects stations, several host plants for Anartia fatima were placed by the researcher ar ound certain nectar sources and other areas that they frequented. Thinking that perhaps the wrong indicator species was chosen to test for abundance (although Morpho peleides sightings were common), the res earcher began to monitor these Anartia host plants in addition to the other experimental pl ants. Despite being within an arms reach of

PAGE 74

74 literally hundreds of Anartia over the course of five days, not a single egg was found on the host plants. After the experiment was complete, the rese archer began to comp are the north central Florida test versus the Costa Ri ca test in order to reason why there was such a drastic difference in oviposition levels. There was indeed one glari ng difference between the trials. In both north central Florida and Costa Rica, orange marking fl ags were used to delineate transects and mark the experimental host plants. It could have b een possible that Monarch females saw the orange marking flag and were attracted to it as bei ng a possible mate, subsequently ovipositing on the plant, or returning at a later time to oviposit. At this point, shiny blue plastic material was used to mimic butterfly presence in th e same way the orange flags were presumed to do. After the blue plastic was tied to the host plants, the 10-da y experiment was repeated (Figure 4-7). Again, there was no oviposition. Although the orange marking flags could have served as an attractant in the north central Florida trials, there is no proof. This brings us to the question of why M onarch butterflies were the only species that oviposited during the north central Florida trials? Although Swa llowtails and Buckeyes were seen regularly in and around the farms, there wa s virtually no oviposition (there were two eggs seen of the Swallowtail during th e entire experiment). There are two main factors believed to be the cause for this lack of oviposition. The first is a relative lack of visual conspicuousness. Courtney (1982) suggested that the level of visu al conspicuousness (i.e., how obvious a plant is compared to its surroundings) directly affects ovi position in butterflies. Specifically, as visual conspicuousness increases, so does oviposition. Ba sed on the size, color, and leaf shape of the two host plants used for the Swallowtail and Buck eye (parsley and plantain, respectively), they likely exhibited a relatively low level of visual conspicuousness in relation to their surroundings.

PAGE 75

75 These plants were low to the ground, had leaves only 10 cm above the grass line, grew almost prostrate in the field, and did not have any flowers or other colo r patterns to distinguish them from the surrounding grasses. Th ese qualities are nearly the exact opposite of the milkweed used for the Monarch butterflies. The milkweeds grew ve rtically, exhibited a hei ght of about meter, commonly had flowering parts, and the leaves were well above the grass line. A second possible reason for the lack of ovipos ition is an improper choice of host plants, despite the fact that these were the host plants actively used for rearing of the respective butterfly species at both farms. Rausher (1979a) sugge sts that landing frequencies among butterflies may fluctuate and change seasonally, depending on host abundance and qua lity. It is possible that there was a temporary or seasonal host plant shif t prior to the start of the experiment. In addition, Courtney and Forsberg (1988) suggest that landing frequencies may depend on the relative abundance of a preferred host plant. Should the numbers or quality of these host plants in the wild have declined recently, or perhaps qua ntities of another host pl ant surged prior to the experiment, landing frequencies, and consequently oviposition, could have been negatively affected for these butterflies.

PAGE 76

76 Figure 4-1. Although most host plants were situated in fields like this at Shady Oaks Butterfly Farm, some plants were in more fore sted areas as seen in the background

PAGE 77

77 Figure 4-2. More open layout of the area surrounding Great house farm. In the distance, there is a residence with a butterfly garden a probable a ttractive factor

PAGE 78

78 Figure 4-3. A residence in the area surrounding Greathouse farm, comp lete with butterfly garden

PAGE 79

79 Figure 4-4. Large screen enclosure housing hundreds of milkweed plants

PAGE 80

80 Figure 4-5. Student groups partic ipating in ecotourism may be perc eived as threatening by larger butterfly predators, such as birds and lizards

PAGE 81

81 Figure 4-6. Heliconia plants around the exterior of the fa rm epicenter could have acted as competition for oviposition in the area

PAGE 82

82 Figure 4-7. Blue plastic material used in hopes of stimulating oviposition by Morpho peleides

PAGE 83

83 CHAPTER 6 CONCLUSIONS The results from this experiment show that there is indeed an effect that the typical butterfly farm has on wild butte rfly oviposition, and in turn, w ild butterfly abundances. The tendency for butterflies to be in greater abundances closer to the farm could be the result of a combination of limited dispersal among butterflies from within the farm, whether escaped or hatched in the nurseries, as well as wild butterfli es being attracted to the farm epicenter and the butterfly resources it contains. It is possible that the farm is emitting specific cues indicating habitat quality as well as the presence of viable mates, although untested in this experiment. This experiment cannot prove or disprove that thes e cues are real, nor if there is indeed a force attracting wild butterflies. However, we can say that there are greater butterfly abundances closer to the farm, whether from wild populations or but terflies from within the farm. It would be a very interesting and helpful follow-up study to examine these cues and the attractive stimuli these farms may elicit with regards to wild butterflies. By testing for larval predation around the farm we see that the farm poses no clear threat of abnormally high levels of predation closer to th e farm. In fact, the oppos ite is the case. There appears to be abnormally low levels of predation as one gets closer to the farm. This could be due to discrepatncies in predator populations and s earching behavior. It could also be the result of hightened levels of human activity, in the form of farm workers and ecotourists, which may be perceived as a threat by potential predators. By testing for egg parasitism, the data suggest that proximity to the farm does not elicit higher levels of parasitism when compared to more distal areas. However, zero parasitism compared with zero parasitism does not provide convincing enough statistic s to author itatively claim no difference.

PAGE 84

84 An initial hypothesis of the researcher was that butterfly farms may be acting as ecological traps, presenting adult butterflies with seemingly suita ble resources and habitat, only to have their eggs parasitized and larvae predated upon at higher than normal rates. Although it was evident that butterflies we re in higher numbers surrounding the farm, larval predation was less of a threat at the farm and especially low quite near the farm epicenter. Therefore, this study suggests that butterfly farms may, in fact, not se rve as ecological traps an d in a way could have the opposite effect by acting as a source for metapopulations in fragmented landscapes. Although this is a relatively unsubstantiated assumption, it could provide the framework for an intriguing follow-up study. Being a preliminary study of butterfly abundance, larval predation, a nd egg parasitism at butterfly farms and ecotourism areas, this set of experiments yielded interesting and helpful results. Many possible explanations for trends in these data involve bu tterfly activity from within the epicenter as well as possible attraction to the epicenter. Further studies should concentrate on the specific areas in question, su ch as nurseries, outdoor gardens, and breeding and rearing centers in order to conclusively ev aluate specific attractive qualities they may possess.

PAGE 85

85 LIST OF REFERENCES ABRAMS, P. A., O. LEIMAR, S. NYLIN, AND C. WIKLUND. 1996. The effect of fl exible growth rates on optimal sizes and development times in a seasonal environment. American Naturalist 147: 381-395. ANDOW, D. A., AND D. R. PROKRYM, D. R. 1990. Plant structural co mplexity and host-finding by a parasitoid. O ecologia 82: 162-165. ARENDT, J. D. 1997. Adaptive intrinsic growth rates: an integration across taxa. The Quarterly Review of Biology 72 (2): 149-177. ATKINSON, D. 1994. Temperature and organism size a biological law for ectotherms? Advances in Ecologica l Research 25: 1-58. ATKINSON D., S. A. MORLEY, D. WEETMAN, AND R. N. HUGHES. 2001. Offspring size responses to maternal temperature in ectotherms. In D. Atkinson and M. Thorndyke (Eds.). Environment and Animal Development: Genes, Life Histories, Plasticity, pp. 269-286. BIOS Scientific Publishers, Oxford, UK. AVELAR, T. 1993. Egg Size in Drosophila: Standard Unit of Investment or Variable Response to Environment? The Effect of Temperature. Journal of Insect Physiology 39 (4): 283289. AZEVEDO, R. B. R., V. FRENCH, AND L. PARTRIDGE. 1996. Thermal evolution of egg size in Drosophila melanogaster Evolution 50: 2338-2345. BAGUETTE, M., I. CONVIE, AND G. NEVE. 1996. Male density affects fe male spatial behavior in the butterfly Proclossiana eunomia. Acta Oecologia 17: 225-232. BAGUETTE, M., C. VANSTEENWEGEN, AND I. CONVI. 1998. Sex-biased density-dependent migration in a metapopulat ion of the butterfly Proclossiana eunomia Acta Oecologia 19: 17-24. BAKER, R. R. 1969. The evolution of the migratory ha bit in butterflies. Journal of Animal Ecology 38: 703-746. BERNAYS, E. A. 1997. Feeding by Lepidopteran larvae is dangerous. Ecological Entomology 22: 121-123. BOPPRE, M. 1983. Leaf-scratching-a speci alized behavior of danain e butterflies (Lepidoptera) for gathering secondary plant substances. Oecologia 59: 414-416. BOWLER, D. E., AND T. G. BENTON. 2005. Causes and consequences of animal dispersal strategies: relating individual behavior to spatial dynamics. Biological Review 80: 205225.

PAGE 86

86 BROMMER, J. E., AND M. S. FRED. 1999. Movement of the Apollo butterfly Parnassius apollo related to host plant and nectar plant patches. Ec ological Entomology 24: 125-131. BROWER, L. P. 1969. Ecological Chemistry. Scientific American. 220: 22-29. ________. 1984. Chemical defense in butterflies. In R. I. Vane-Wright and P. R. Ackery (Eds.) The Biology of Butterflies, pp. 109-134. Academic Press, London. ________, AND L. S. FINK. 1985. A natural toxic defense system: cardenolides in butterflies vs. birds. In N. S. Braveman and P. Bronstein (Eds.) Experimental Assessments and Clinical Applications of Conditione d Food Aversions, pp. 171-188. New York Academy of Sciences, New York. CLOBERT, J., E. CANCHIN, A. A. DHONDT, AND J. D. NICHOLS. 2001. Dispersal. Oxford University Press, Oxford, UK. COHEN, J. A., AND L. P. BROWER. 1982. Oviposition and larval success of wild monarch butterflies (Lepidoptera: Danaidae) in re lation to host plant size and cardenolide concentration. Journal of the Kans as Entomological Society 55: 343-348. COURTNEY, S. P. 1982. Coevolution of pierid butterflies and their cruciferous foodplants. IV. Host apparency and Anthocharis cardamines oviposition. Oecologia 52: 258-265. ________, AND J. FORSBERG. 1988. Host use by two pierid butterf lies varies with host density. Functional Ecology 2: 67-75. ________, G. K. CHEN, AND A. GARDNER. 1989. A general model for individual host selection. Oikos 55:55-65. ________, AND T. T. KIBOTA. 1990. Mother doesnt know best: selection of hosts by ovipositing insects. In E. A. Bernays (Ed.) Insect-Plant In teractions, pp. 161-188. CRC Press, Boca Raton, Florida. CRAIG, T. P., AND T. OHGUSHI. 2002. Preference and performance are correlated in the spittlebug Aphrophora pectoralis on four species of willow. Ecological Entomology 27: 529-540. DEMPSTER, J. P. 1983. The natural control of populations of butterflie s and moths. Biological Reviews 58: 461-481. DETHIER, V. G. AND R. H. MACARTHUR. 1964. A fields capacity to support a butterfly population. Nature 201: 728-729. DIXON, C. A., J. M. ERICKSON, D. N. KELLETT, M. ROTHSCHILD. 1978. Some adaptations between Danaus plexippus and its food plan t, with notes on Danaus chrysippus and Euploea core (Insecta: Lepidoptera). Journal of Zoology 185 (4): 437-467.

PAGE 87

87 DUNNING, J. B. J., D. J. STEWART, B. J. DANIELSON, B. R. NOON, T. L. ROOT, R. H. LAMBERSON, AND E. E. STEVENS. 1995. Spatially explicit populati on models: current forms and future uses. Ecological Applications 5: 3-11. DUSSOURD, D. E. 1990. The vein drain; or how insects ou tsmart plants. Natural History 90: 4449. EHRLICH, P. R., R. R. WHITE, M. C. SINGER, S. W. MCKECHNIE, AND L. E. GILBERT. 1975. Checkerspot butterflies: A histori cal perspective. Science 188: 221-228. ERNSTING, G., AND J. A. ISAAKS. 1997. Effects of temperature and season on egg size, hatchling size and adult size in Notiophilus biguttatus. Ecological Entomology 22 (1): 32-40. FATOUROS, N. E., G. B. KISS, L. A. KALKERS, R. S. GAMBORENA, M. DICKE, AND M. HILKER. 2005. Oviposition-induces plant cues: do they arrest Trichogramma wasps during host location? Entomologia Experimentalis et Applicata 115: 207-215. FEENY, P. 1991. Chemical constraints on the evol ution of swallowtail butterflies. In J. Wiley (Ed.) Plant-Animal Inter actions: Evolutionarey Ecology in Tropical and Temperate Regions, pp. 315-340. New York. FISCHER K., P. M. BRAKEFIELD AND B. J. ZWAAN. 2003. Plasticity in butterfly egg size: why larger offspring at lower te mperatures? Ecology 84: 3138-3147. ________, A. N. M. BOT, P. M. BRAKEFIELD, AND B. J. ZWAAN. 2006a. Do mothers producing large offspring have to sacrifice fecundity ? Journal of Evolutionary Biology 19: 380391. ________, S. S. BAUERFEIND, AND K. FIEDLER K. 2006b. Temperature-mediat ed plasticity in egg and body size in egg size-selected lines of a butterfly. Journal of Thermal Biology 31: 347-354. FORISTER, M. L. 2004. Oviposition preference and larval performance within a diverging lineage of lycaenid butterflies. Ecol ogical Entomology 29: 264-272. FOX, C. W., AND M. E. CZESAK. 2000. Evolutionary ecology of progeny size in arthropods. Annual Review of Entomology 45: 341-369. FOX, R. M. 1966. Forelegs of butterflies I. Introduction: chemoreception. Journal of Research on the Lepidoptera 5: 1-12. GATES, J. E. AND L. W. GYSEL. 1978. Avian nest dispersion and fledging success in field-forest ecotones. Ecology 59: 871-883. GILBERT, I. E. AND M. C. SINGER. 1973. Dispersal and gene flow in a butterfly species. American Naturalist 107: 58-72.

PAGE 88

88 ________, AND ________. 1975. Butterfly ecology. A nnual Review of Ecology and Systematics. 6: 465-397. GINGRAS, D., AND G. BOIVIN. 2002. Effect of plant structure, host density and foraging duration on host finding by Trichogramma evanescens (Hymoptera: Trichogrammatidae). Environmental Entomology 31: 1153-1157. GLENDINNING, J. I., AND L. P. BROWER. 1990. Feeding and breeding responses of five mice species to overwintering aggregations of the monarch butterfly. Journal of Animal Ecology 59: 1091-1112. GORDON, I., AND W. AYIEMBA. 2003. Harnessing butterfly biodiversity for improving livelihoods and forest conservation: The Ki pepeo Project. Journal of Environment & Development 12 (1): 82-98. GOTTHARD, K. 1999. Life history analysis of growth strategies in temperate butterflies. Ph.D. Thesis, Stockholm University. ________. 2000. Increased risk of predation as a cost of high growth rate: an experimental test in a butterfly. Journal of Animal Ecology 69: 896-902. HANSKI, I. AND M. E. GILPIN. 1997. Metapopulation Biology: Ecology, Genetics and Evolution, Academic Press, San Diego. ________. 1999. Metapopulation Ecology, Oxford University Press, Oxford, UK. HARRIS, M. O., M. SANDANYAKA, AND W. GRIFFIN. 2001. Oviposition preferences of the Hessian fly and their consequences for the survival and reproductive potential of offspring. Ecological Entomology 26: 473-486. HOVANITZ, W., AND V. C. S. CHANG. 1963. Change of food-plant preference by larvae of Pieris rapae controlled by strain selection, and the in heritance of this trait. Journal of Research on the Lepidoptera 1: 163-168. ILSE, D. 1937. New observation on responses to colo rs in egg-laying butterflies. Nature 140: 544-545. JAMES, M., F. GILBERT, AND S. ZALAT. 2003. Thyme and isolation for the Sinai baton blue butterfly ( Pseudophilotes sinaicus ). Oecologia 134: 445-453. JANZEN, D. H. 1967. Synchronization of sexual reproduc tion of trees with the dry season in Central America. Evolution 21: 620-637. KLOMP, H., B. J. TEERINK. 1962. Host Selection and Number of Eggs per Oviposition in the EggParasite Trichogramma embr yophagum. Nature 195: 1020-1021.

PAGE 89

89 KUUSSAARI, M., M. NIEMINEN, AND I. HANSKI. 1996. An experimental study of migration in the Glanville fritillary butterfly Melitaea cinxia Journal of Animal Ecology 65: 791-801. _________, M., I. SACHERI, AND M. CAMARA. 1998. Allee effect and popoulation dynamics in the Glanville fritillary bu tterfly Oikos 82: 384-392. KYI, A., M. P. ZALUCKI, AND I. J. TITMARSH. 1991. An experimental study of early stage survival of Helicoverpa armigera (Lepidoptera: Noctuidae) on cotton. Bulletin of Entomological Research 81: 263-271. LEVINS, R. 1968. Evolution in Changing Environments Princeton University Press, New Jersey. _________, AND R. H. MACARTHUR. 1969. An hypothesis to explain the incidene of monophagy. Ecology 50: 910-911. LI, Y. L. 1994. Worldwide use of Trichogramma for biological control of different crops: a survey. In E. Wajnberg, and S. A. Hassan (E ds.). Biological control with egg parasitoids pp. 37-53. Berkshire: CAB International. LUCANSKY, T. W., AND K. T. CLOUGH. 1986. Comparative anatomy and morphology of Asclepias perennis and Asclepias tuberosa subspecies rolfsii Botanical Gazette 147: 290-301. LUKIANCHUK, J. L., AND S. M. SMITH. 1997. Influence of structural complexity on the foraging success of Trichogramma minutum : a comparison of search on artificial and foliage models. Entomolgia Experimentlis et Applicata 84: 221-228. MALCOLM, S. B., AND L. P. BROWER. 1989. Evolutionary and eco logical implications of cardenolide sequestration in the monarc h butterfly. Expe rientia 45: 284-295. MAYHEW, P. J. 1997. Adaptive patterns of host-plant selection by phytophagous insects. Oikos 79: 417-248. MYERS, J. H. 1985. Effects of physiological condition of the host plant on the ovipositional choice of the cabbage white butterfly, Pieris rapae Journal of Animal Ecology 54: 193-204. NYLIN, S., AND K. GOTTHARD. 1998. Plasticity in the life history traits. Annual Review of Entomology 43: 63-83. ODENDAAL, F. J., Y. IWASA, AND P. R. EHRLICH. 1985. Duration of female availability and its effect on butterfly mating systems. American Naturalist 125: 673-678. ________, P. TURCHING, AND F. R. STERMITZ. 1989. Influence of host-plant densit and male harassment on the distribution of female Euphydras anicia (Nymphalidae). Oecologia 78: 283-288.

PAGE 90

90 OMENGE, P. I. 2002. The Role of Butterfly Farming in Forest Conservation and Community Development in Kenya. Swedish University of Agricultural Sciences. Masters Thesis No.16 PAK, G. A., I. VAN HALDER, R. LINDEBOOM, AND J. J. G. STROET. 1985. Ovarian egg supply, female age and plant spacing as factors influencing searching activity in the egg parasited Trichogramma sp. Medelelingen van de Faculteit Landbouwwetenshappen Rijksuniversiteit Gent 50 (2a) 369-378. PINTO, J. D., AND R STOUTHAMER. 1994. Systematics of the Trichogrammatidae with emphasis on Trichogramma In E. Wajnberg, and S. A. Hassan (Eds.). Biological Control with Egg Parasitoids pp. 1-36. Berkshire: CAB International. PROKOPY, R. J., AND E. D. OWENS. 1983. Visual detection of plan ts by herbivorous insects. Annual review of Entomology 28: 40-48. RABASA, S. G., D. GUITIERREZ, AND A. ESCUERDO. 2005. Egg laying by a butterfly on a fragmented host plant: a multi-level approach. Ecography 28: 629-639. RAUSHER, M. D. 1979a. Larval habitat suitability a nd oviposition preference in three related butterflies. Ecol ogy 60 (3): 503-511 ________. 1979b. Egg recognition: its a dvantage to a butterfly. Animal Behavior 27: 1034-1040. RAY, C., M. GILPIN, AND A. T. SMITH. 1991. The effect of conspecific attraction on metapopulation dynamics. Biological Jounr al of the Linnaean Society 42: 123-134. RENWICK, J. A. A., AND C. D. RADKE. 1980. An oviposition deterrent as sociated with frass from feeding larvae of the cabbage looper. Trichopulsia ni (Lepidoptera: Noctuidae). Environmental Entomology 9: 318-320. ________, AND _______. 1982. Activity of cabbage extracts in tererring oviposition by the cabbage looper, Trihoplusia ni. Proceedings of the Fifth In ternational Symposium of Insect-Plant Relationships Wageningen: Pudoc. pp. 139-43. ROMEIS, J., D. BABENDREIER, F. L. WACKERS, AND T. G. SHANOWER. 2005. Habitat and plant specificity of Trichogramma egg parasitoids underlying mechanisms and implications. Basic and Applied Ecology 6: 215-236. ROTHSCHILD, M., AND L. M. SCHOONHOVEN. 1977. Assessment of egg load by Pieris brassicae (Lepidoptera: Pieridae). Nature 266: 352-355. SCHLAEPFER, M. A., M. C. RUNGE, AND P. W. SHERMAN. 2002. Ecological and evolutionary traps. Trends in Ecology and Evolution. 17 (10): 474-480.

PAGE 91

91 SCHURR, K., AND F. G. HOLDAWAY. 1970. Olfactory responses of female Ostrinia nubilalis (Lepidoptera: Pyraustinae). Entomologia E xperimentalis et Applicata 13: 455-461. SHAPIRO, A. M. 1970. The role of sexual behavior in density-related disperal of pierid butterflies. American Naturalist 104: 367-372. ________, AND R. T. CARDE. 1970. Habitat selection and comp etition among sibling species of satyrid butterflies. Evolution 24: 48-54. SHREEVE, T. G. 1995. Butterfly mobility. In A. S. Pullin (Ed.). Ecology and conservation of butterflies, pp. 37-45. Chapman and Hall, London. SINGER, M. C. 1993. Behavioral constrains on the evolut ionary expansion of an insect diet: a case history. In L. Real (Ed.) Behavioral Mechanisms in Evolutionary Ecology Chicago: University of Chicago press. SMITH, C. C., AND FRETWELL. 1974. The optimal balance between size and number of offspring. American Naturalist 108: 499-506. SMITH, S. M. 1996. Biological control with Trichogramma : advances, successes, and potential of their use. Annual Review of Entomology 41: 375-406. STANTON, M. L. 1984. Short-term learning and the search ing accuracy of egg-laying butterflies. Animal Behavior 32: 33-40. STEARNS, S. C. 1992. The Evolution of Life Histories 1st edition, Oxford University Press, Oxford, UK. STEINGENGA M. J., AND K. FISCHER. 2007. Ovarian dynamics, egg size, and egg number in relation to temperature and mati ng status in a butterfly. En tomologia Experimentalis et Applicata 125: 195-203. SURVEKROPP, B. P. 1997. Host-finding behavior of Trichogramma brassicae in maize. Ph.D. Dissertation, Wageningen Agricultural University, Wageningen, The Netherlands. THOMSON, J. N. 1988. Variation in preference and sp ecificity in monophagous and oligophagous swallowtail butterflies. Evolution 42: 118-128. TSCHARNTKE, T., AND R. BRANDAL. 2004. Plant-insect interactions in fragmented landscapes. Annual Review of Entomology 49: 405-430. WAJNBERG, E. 1994. Intra-population genetic variation in Trichogramma In E. Wajnberg, and S. A. Hassan (Eds.). Biological control with egg parasitoids pp. 245-271. Berkshire: CAB International.

PAGE 92

92 WERNER, E. E., AND B. R. ANHOLT. 1993. Ecological consequences of the trade-off between growth and mortality rates mediated by foraging activity. American Naturalist 142: 242-272. WIKLUND, C. 1977. Oviposition, feeding and spatia l separation of breeding and foraging habitats in a population of Leptidea sinapis (Lepidoptera). Oikos 28: 56-68. WILLIAMS, B. K., AND J. D. NICHOLS. 1984. Optimal timing in biological processes. American Naturalist 123: 1-19. WOLFSON, J. L. 1980. Oviposition response of Pieris rapae to environmentally induced variation in Brassica nigra Entomologia Experimentalis et Applicata Entomologia Experimentalis et Applicata 27: 223-232. YAMPOLSKI L. Y., AND S. M. SCHEINER. 1996. Why larger offspring at lower temperatures? A demographic approach. Am erican Naturalist 147: 86-100. ZALUCKI, M. P., AND R. L. KITCHING. 1982. Temporal and spatial vari ation of morta lity in field populations of Danaus plexippus L. and D. chrysippus L. larvae (Lepidoptera: Nymphalidae). Oecologia 3: 201-207. ________, S. B. MALCOLM, T. D. PAINE, C. C. HANLON, L. P. BROWER, AND A. R. CLARKE. 2001. Its the first bites that count: Survival of first-instar monarchs on milkweeds. Austral Ecology 26 (5): 547.

PAGE 93

93 BIOGRAPHICAL SKETCH Court W helan was born in July of 1983, in Clearwater, Florida. His love for nature and a passion for living things began al most immediately and progressed through his par ticipation in nature programs, from summer camps to school fi eld trips. His formal training in biology and entomology began during the second semester of his undergraduate career at the University of Florida (Spring 2002), upon taking a basic ento mology course taught by Dr. Don Hall of the department of Entomology and Nematology. After m eeting with Dr. Hall to discuss his affinity for entomology and the programs and advanced cla sses that were offered at UF, Court declared his major as entomology immediately and has never looked back since. During his junior year of undergraduate work (2004), he was offered a studen t internship with an ecotourism tour operator to co-lead a group of 18 high school students on a 2-week field trip to the rainforests and coral reefs of Belize. With the responsibility to instruct the students on tropical entomology and ecology as well as co-lead the group throughout the trip, a strong passion for ecotourism emerged simultaneously with a passion for trop ical entomology and ecology. Following the trip, his passion for ecotourism and entomology only grew stronger, as he immediately began working with Dr. Thomas Emmel, Dr. Jaret Da niels and Dr. John Capine ra on developing a new graduate program entitled ecotourism entomology. While continuing his undergraduate career, he began working with a local tour operator, which specialized in tropical ecotourism, in arranging and leading ecotourism trips. After receiving his Bach elor of Science degree in the Spring of 2005, Court received an Alumni Fellowshi p from the University of Florida and was admitted to graduate school in the Department of Entomology and Nematology. He received his M.S. degree in 2008 and continue s to arrange and lead ecotour ism trips that are focused on tropical biology, ecology and entomology while pursuing his Ph.D. in the newly formed discipline of ecotourism entomology.































Figure 4-4. Large screen enclosure housing hundreds of milktweed plants









patch's carrying capacity (Dethier and MacArthur 1964). As the butterflies deplete the host and

nectar plant resources in the habitat, a "migratory threshold" would be reached (Baker 1984).

This threshold is a theoretical limit to the rate of migration so that as the threshold is exceeded,

dispersal away from the patch would begin to surpass dispersal into the patch, continuing this

way until returning to a stable density. This stable density would be below the carrying capacity

and allow key resources of the habitat such as nectar and host plants to regenerate, allowing for

more immigration in the future.

Dispersal and the Farms

The notion of conspecific density is of particular importance to this study of butterfly

farms, which have artificially high conspecific densities. It is clear that farm enclosures are

artificially stocked with butterflies, yielding this artificially high conspecific density. However,

it has not been documented whether or not butterflies occur in high numbers around the exterior

of the farm epicenter.

The enclosures at butterfly farms present a special circumstance when considering habitat

patches and conspecific densities. They may be treated as separate habitats due to the walls

excluding immigration and migration. At the same time, however, they may still serve as an

attractive force (i.e., presence of host plants, nectar sources, pheromones and visual signals

emanating from potential mates, etc.) to wild populations of butterflies, in which case they could

be considered part of a larger patch that includes all areas of the butterfly farming operation.

According to Hanski et al. (1994) and Kuussaari et al. (1998), butterflies may be attracted to

high conspecific densities in search of an appropriate mate. Additionally, Gilbert and Singer

(1973) suggested that high conspecific densities may also attract butterflies when acting as a cue

for good habitat quality. Should these theories apply to butterfly farms as well, a great problem













3.0

2.5

2.0

1.5

1.0

0.5


1 2 3 4 5 6 7 8 9 10 11
SEgg Number 3.033 2.933 2.933 2.700 1.683 1.283 1.183 1.567 1.067 1.400 1.633
Transect (at 10m intervals)


Figure 3-1. Summary graph of average egg number for each transect

Table 3-1. Summary statistics for North central Florida oviposition experiment
Sample Mean egg
Transect size number St Dev Qmax Q3 Median Q1 Min
1 60 3.033 3.835 15 6 1 0 0
2 60 2.933 3.901 14 4.5 1 0 0
3 60 2.933 4.137 14 4 1 0 0
4 60 2.700 4.777 21 3 0 0 0
5 60 1.683 2.861 13 2 0 0 0
6 60 1.283 2.387 10 1 0 0 0
7 60 1.183 1.987 9 2 0 0 0
8 60 1.567 3.061 12 2 0 0 0
9 60 1.067 1.812 7 2 0 0 0
10 60 1.400 2.688 11 2 0 0 0
11 60 1.633 3.650 17 1.5 0 0 0


Table 3-2. Si nfcance values obtained from the SAS aroc mixed procedure (oc=0.1)

32.18 <.0001
1.42 0.2049
13.64 <.0001


24.25


<.0001










true, it is a unique example of the benefits of integrating ecotourism (especially the high energy

and noisy student groups) into a butterfly farming operation.

Although statistical T-tests indicate that Costa Rica did not follow the same pattern of

significant differences for north central Florida, the raw data clearly follow the same pattern as

north central Florida. For this reason, the researcher does not feel that there is any notable

difference between the Costa Rica and north central Florida trials. In order to test for significant

differences between the localities, additional studies are needed.

There are numerous possibilities for why an overall decline in predation was observed as

one gets closer to the farm epicenter, but we cannot ignore the plain fact that there were

statistically significant differences allowing us to make this conclusion in the first place. There

is a great need for further research to be done on predation rates in regards to proximity to

butterfly farms. In particular, future studies should concentrate on causes for these unexpected

results.

Egg Parasitism in Florida and Costa Rica

Egg parasitism was found to be non-existent among monarch butterfly eggs during the

10-day experiment at all three geographic locations in Florida and Costa Rica. Although this

bodes well for the prognosis of overall parasitism rates around these particular butterfly farms,

we cannot make any general conclusions from this experiment. These results certainly do not

mean that parasites and parasitoids are absent at butterfly farms, and they likely do not mean that

they are necessarily absent from the farms used in this experiment. What is likely is that the

experiment was not designed in a way to "attract" parasites and parasitoids to the butterfly eggs.

The first problem is that the butterfly eggs may not have been available to the parasitoids

for long enough. In a study by Fatouros et al. (2005), wasps were found to be more attracted to

leaves with eggs deposited for 48 hr, rather than eggs for 24 hr. Fatouros et al. (2005) also found









Butterfly Abundance and Oviposition at the Northern Costa Rica Farm

The results from the butterfly abundance study in Costa Rica were unexpected and very

surprising. With the number of eggs deposited in the north central Florida trials ranging from

several hundred to almost one thousand during the 10-day experiment, finding absolutely zero

eggs during the Costa Rica trial is perhaps as interesting as if there were hundreds of eggs found.

There are several factors that may contribute to this overall lack of oviposition.

One explanation has to do with the extreme competition among host plants in the tropical

environment. As stated before, butterflies make tremendous use of visual stimuli to locate host

plants, particularly the color and shape of the host plants' leaves (Stanton 1984). Furthermore,

relative abundance of host plants in an area has been shown to affect landing frequencies by

gravid female butterflies (Courtney & Forsberg 1988). Consequently, the tremendous diversity

of plants in the rainforest, some as host plants for these species, others with similarly shaped and

colored leaves, could have acted as competition towards the potted plants placed out in this

experiment. With the experiment lasting 10 days, perhaps this was too short of a duration for

butterflies to recognize and learn the location of such host plants in a sea of both nearly identical

and intricately complex stimuli.

Seasonality may have played a part in the visual homogeneity of the landscape as well.

This experiment was conducted during the rainy season in Costa Rica and, thus, there was little

to no flowering exhibited by the plants. According to Janzen (1967), tropical plants in

Guanacaste, Costa Rica, flower and fruit in the dry season due to the fact that investing into non-

vegetative activity is least detrimental to their vegetative competitive ability when rains are less

frequent. The presence of flowers may have increased oviposition by providing another visual

cue for the butterflies, which may have yielded oviposition on the host plant.