Citation
Senescence and Other Factors Affect Fecundity in Two Species of Culex Mosquitoes

Material Information

Title:
Senescence and Other Factors Affect Fecundity in Two Species of Culex Mosquitoes
Creator:
MCCANN, SEAN MICHAEL ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Blood ( jstor )
Body size ( jstor )
Eggs ( jstor )
Fecundity ( jstor )
Female animals ( jstor )
Linear regression ( jstor )
Modeling ( jstor )
Rafts ( jstor )
Regression analysis ( jstor )
Species ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Sean Michael McCann. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
3/1/2007
Resource Identifier:
649810219 ( OCLC )

Downloads

This item has the following downloads:

mccann_s ( .pdf )

mccann_s_Page_005.txt

mccann_s_Page_079.txt

mccann_s_Page_093.txt

mccann_s_Page_024.txt

mccann_s_Page_057.txt

mccann_s_Page_099.txt

mccann_s_Page_004.txt

mccann_s_Page_096.txt

mccann_s_Page_071.txt

mccann_s_Page_097.txt

mccann_s_Page_049.txt

mccann_s_Page_104.txt

mccann_s_Page_086.txt

mccann_s_Page_098.txt

mccann_s_Page_019.txt

mccann_s_Page_110.txt

mccann_s_Page_080.txt

mccann_s_Page_102.txt

mccann_s_Page_018.txt

mccann_s_Page_051.txt

mccann_s_Page_088.txt

mccann_s_Page_075.txt

mccann_s_Page_060.txt

mccann_s_Page_047.txt

mccann_s_Page_054.txt

mccann_s_Page_085.txt

mccann_s_Page_077.txt

mccann_s_Page_045.txt

mccann_s_Page_052.txt

mccann_s_Page_066.txt

mccann_s_pdf.txt

mccann_s_Page_006.txt

mccann_s_Page_084.txt

mccann_s_Page_017.txt

mccann_s_Page_009.txt

mccann_s_Page_035.txt

mccann_s_Page_053.txt

mccann_s_Page_021.txt

mccann_s_Page_037.txt

mccann_s_Page_003.txt

mccann_s_Page_065.txt

mccann_s_Page_031.txt

mccann_s_Page_042.txt

mccann_s_Page_036.txt

mccann_s_Page_108.txt

mccann_s_Page_091.txt

mccann_s_Page_070.txt

mccann_s_Page_025.txt

mccann_s_Page_103.txt

mccann_s_Page_034.txt

mccann_s_Page_105.txt

mccann_s_Page_011.txt

mccann_s_Page_026.txt

mccann_s_Page_068.txt

mccann_s_Page_106.txt

mccann_s_Page_039.txt

mccann_s_Page_101.txt

mccann_s_Page_038.txt

mccann_s_Page_013.txt

mccann_s_Page_083.txt

mccann_s_Page_001.txt

mccann_s_Page_073.txt

mccann_s_Page_094.txt

mccann_s_Page_092.txt

mccann_s_Page_022.txt

mccann_s_Page_089.txt

mccann_s_Page_063.txt

mccann_s_Page_081.txt

mccann_s_Page_072.txt

mccann_s_Page_050.txt

mccann_s_Page_002.txt

mccann_s_Page_007.txt

mccann_s_Page_014.txt

mccann_s_Page_027.txt

mccann_s_Page_058.txt

mccann_s_Page_012.txt

mccann_s_Page_030.txt

mccann_s_Page_008.txt

mccann_s_Page_044.txt

mccann_s_Page_023.txt

mccann_s_Page_015.txt

mccann_s_Page_059.txt

mccann_s_Page_056.txt

mccann_s_Page_067.txt

mccann_s_Page_020.txt

mccann_s_Page_074.txt

mccann_s_Page_029.txt

mccann_s_Page_041.txt

mccann_s_Page_010.txt

mccann_s_Page_109.txt

mccann_s_Page_061.txt

mccann_s_Page_033.txt

mccann_s_Page_090.txt

mccann_s_Page_043.txt

mccann_s_Page_055.txt

mccann_s_Page_107.txt

mccann_s_Page_082.txt

mccann_s_Page_069.txt

mccann_s_Page_040.txt

mccann_s_Page_046.txt

mccann_s_Page_064.txt

mccann_s_Page_076.txt

mccann_s_Page_028.txt

mccann_s_Page_078.txt

mccann_s_Page_062.txt

mccann_s_Page_095.txt

mccann_s_Page_100.txt

mccann_s_Page_087.txt

mccann_s_Page_048.txt

mccann_s_Page_016.txt

mccann_s_Page_032.txt


Full Text





SENESCENCE AND OTHER FACTORS AFFECT FECUNDITY INT TWO SPECIES OF
Cudex MOSQUITOES (DIPTERA: CULICIDAE)




















By

SEAN MICHAEL MCCANN


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

2006

































Copyright 2006

by

Sean Michael McCann




































To my parents.









ACKNOWLEDGMENTS

I gratefully acknowledge the assistance of my supervisory committee, the laboratory staff,

and various faculty around FMEL for their assistance with time, material or advice.












TABLE OF CONTENTS


page

ACKNOWLEDGMENT S .............. ...............4.....


LI ST OF T ABLE S ............ ....._._. ...............8.....


LI ST OF FIGURE S .............. ...............9.....


AB S TRAC T ............._. .......... ..............._ 1 1..


CHAPTER


1 INTRODUCTION .............. ...............13....


Mosquitoes............... ...............1
Taxonomy ................. ............. ... ........ ...... .........13
Distribution and Notes on Natural History ................. ...............13...............
Vector Associations ................. ...............14.................

A ge .............. ...............15....
Bloodmeal Size ................. ...............17.................
Bloodmeal Source............... ...............17.

Body Size and Teneral Reserves .............. ...............18....
M multiple Factors ................. ...............19.................
Sum m ary ................. ...............19.......... ......


2 EFFECTS OF BLOODMEAL SIZE AND BODY SIZE ON FECUNDITY OF WILD
Culex nigripalpus............... ..............2


Introducti on ................. ...............23.................
Materials and Methods .............. ...............24....

Trapping .............. ... ......................2
Hematin Collection and Analysis............... ...............24
Oviposition .............. ...............25....
W i ngl ength s................. ...............25.___......
A nalyses .............. ...............26....
Re sults............. ...... ._ ...............27...
Path Analy si s ............. ...... ._ ...............27...
Discussion ............. ...... ._ ...............28...


3 INFLUENCE OF BLOODMEAL SIZE AND BODY SIZE ON THE FECUNDITY OF
CAPTIVE Culex quinquefasciatus ............ ...... ..._. ...............35....


Introducti on ........._._... ....__.. ...............35....
Materials and Methods .............. ...............36....
Larval Rearing ........._..... ...._... ...............36.....
Pupation and Bloodfeeding .............. ...............36....













Hematin Collection and Quantifieation............... .............3

Oviposition .............. ...............37....
W i ngl ength s............. ...... ._ ...............38...
A nalyses .............. ...............3 8....
Re sults............. ...... ...............39....
Discussion ............. ...... ._ ...............40....


4 Culex nigripalpus AGINTG AND FECUNDITY............... ...............4


Introducti on ................. ...............45.................
Materials and Methods .............. ...............46....

Larval Rearing ................. ...............46.................

Pup ati on ................. ...............46........... ....
Blood feeding .............. ...............47....
Termination of Study ............... ...............47....
Hematin Collection and Quantifieation............... .............4

Oviposition .............. ...............48....
W i ngl ength s............. ...... ._ ...............49...
A nalyses .............. ...............49....
R e sults............. ...... ...............50....
Discussion ............. ...... ._ ...............52....
Conclusions............... ..............5


5 SENESCENCE AND FECUNDITY OF Culex quinquefasciatus ............... ...................6


Introducti on ........._._... ....__.. ...............65....
Materials and Methods .............. ...............67....

Larval Rearing ........._.._.. ...._... ...............67....

Pup ati on ........._..... ...._... ...............67....
Bloodfeeding ................ ..... ...... .... ..........6
Hematin Collection and Quantifieation............... .............6

Oviposition .............. ...............69....
W i ngl ength s............. ...... ._ ...............69...
A nalyses .............. ...............70....
R e sults............. ...... ...............71....
Discussion ............. ...... ._ ...............72....


6 DI SCU SSION ............. ...... ._ ...............83....


Larval Nutrition .............. ...............83....
BI ood meal Size ........._. ....... .__ ............... 4.....

Age ........._ ....... ............... 5.....
Interactions .............. ...............87....
Future Research .............. ...............8 8....













APPENDIX


A WING MEASUREMENT VALIDATION .............. ...............91....


Obj ective ................. ...............91.................
M ethods .............. ...............91....
Re sults ................ ...............91.................
Conclusion ................ ...............91.................


B HEMATIN STANDARD CURVE ................. ...............94.......... .....


Introducti on ................. ...............94.................

Obj ective ................. ...............94.................
M ethod s .............. ...............94....
Re sults ................ ...............95.................
Conclusion ................ ...............95.................


C MOSQUITO WINGLENGTH VS. WEIGHT REGRESSIONS ................ ..................9


Introducti on ................. ...............98.................

Obj ective ................. ........... ...............98.......
Materials and Methods .............. ...............98....
Conclusion ................ ...............99.................


LIST OF REFERENCE S ................. ...............102................


BIOGRAPHICAL SKETCH ................. ...............110......... ......










LIST OF TABLES


Table page

2-1 Summary statistics for bloodmeal size, body size and fecundity of wild Culex
nigripalpus used in analyses. ............. ...............30.....

2-2 Summary of four regressions testing relationships between hematin excreted,
winglength and fecundity. Significance values are for the t-statistic. ............. ................31

2-3 Decomposition of effects in path analysis. Note that the direct effect of bloodmeal
size is the portion of its contribution not attributable to body size. ................. ........._.....34

3-1 Summary statistics for data measured in regression analyses. ............. .....................4

3-2 Summary of four regressions testing relationships between hematin excreted,
winglength and fecundity. ............ ...............42.....

4-1 Summary statistics for controlled and measured parameters used in regression
analyses. .............. ...............58....

4-2 Summary of 5 regression calculations of various factors and combinations of factors
on fecundity. ............. ...............59.....

4-3 Summary of two ANCOVA analyses describing slope of the fecundity versus
standardized hematin curve in 5 age classes............... ...............64

5-1 Summary statistics for factors and responses used in regression models. ................... ......76

5-2 Summaries of five separate regression equations predicting fecundity of Culex
quinquefasciatus fecundity from individual factors and combinations of factors. ............77

5-3 ANCOVA analysis describing slope of the fecundity versus winglength relationship
for 6 Culex quinquefasciatus age classes............... ...............81

A-1 Results of linear regression analysis of callipered length against length measured in
SigmaScan............... ...............9

B-1. Results of the regression analysis for absorbance versus concentration. The slope
estimate was used to calculate all hematin quantities in this work. ............. ..............97

C-1. Results of two regression analyses predicting female mass from winglength. All
regressions were significant at the a=0.05 level. ............. ...............99.....










LIST OF FIGURES


Figure page

1-1. A. Culex nigripalpus Theobald. B. Culex quinquefasciatus Say. ............. ....................21

1-2. Egg raft of Culex quinquefasciatus. These eggs are freshly laid, and have not
darkened. .............. ...............22....

2-1. Example of a wing photograph used to measure winglengths of female Culex
nigripalpus. ........... ...............30......

2-2. Scatterplot of fecundity versus unstandardized hematin for wild Culex nigripalpus.
The line represents the least-squares linear regression. .................. ................3

2-3. Scatterplot of fecundity versus unstandardized winglength for wild Culex nigripalpus.
The line represents the least-squares linear regression. .................. ................3

2-4. Path diagram describing direct effects between the bloodmeal size, body size and
fecundity. ............. ...............34.....

3-1. Scatterplot of fecundity versus bloodmeal size for Culex quinquefasciatus. Data are
presented in unstandardized format. Least squares regression line shown..............._.. ...43

3-2. Scatterplot of fecundity versus winglength for Culex quinquefasciatus. Data are
presented in unstandardized format. Least squares regression line shown. ....................44

4-1. Scatterplot with least squares regression line of the effect of age on fecundity of Culex
nigripalpus. ............. ...............60.....

4-2. Scatterplot with least squares regression line of the effect of body size on fecundity of
Culex nigripalpus. ................. ...............61.__._......

4-3. Scatterplot with least squares regression line of the effect of bloodmeal size on
fecundity of Culex nigripalpus. ............. ...............62.....

4-4. Scatterplot with least squares regression lines of the effect of bloodmeal size on
fecundity of Culex nigripalpus. ............. ...............63.....

5-1. Scatterplot of fecundity versus unstandardized age at bloodfeeding for Culex
quinquefasciatus ................. ...............78.................

5-2. Scatterplot of fecundity versus unstandardized bloodmeal size for Culex
quinquefasciatus. ............. ...............79.....

5-3. Scatterplot of fecundity versus unstandardized body size for Culex quinquefasciatus. .......80

5-4. Scatterplot of unstandardized winglength versus fecundity showing the response
between ages 5- 13 days (solid line, circles) and 17-25 days (broken line, triangles).......81











5-5. Scatterplot of Percentage hatch versus age for Culex quinquefasciatus. ............. .............82

A-1. Regression plot showing relationship between caliper measurement and SigmaScan
measurement. ............. ...............92.....

B-1. Example of a spectral scan of Culex quinquefasciatus excrement. Note peak
corresponding to maximal absorbance for hematin at 387nm. ................... ...............9

B-2. Standard curve used for calculating amount of hematin in mosquito excreta. R2 for this
regression was 0.9969, F=12830 on 1 and 40 degrees of freedom. N=42.. .................. .....97

C-1. Linear regression of winglength on wet mass of pupae of female Culex
quinquefasciatus n=42 ..........._ ..... ..__ ...............100...

C-2. Linear regression of winglength on wet mass of pupae of female Culex nigripalpus.
N =43 .............. ...............101....









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

SENESCENCE AND OTHER FACTORS AFFECT FECUNDITY INT TWO SPECIES OF
Culex MOSQUITOES (DIPTERA: CULICIDAE)

By

Sean Michael McCann

December 2006

Chair: Cynthia C. Lord
Major Department: Entomology and Nematology

Mosquitoes of the genus Culex are considered important vectors of arboviral diseases in

the State of Florida. I investigated the effects of age, body size and bloodmeal size on the

fecundity of Culex nigripalpus and Culex quinquefasciatus using multiple regression models.

I found that in both species, body size and bloodmeal size were strong predictors of

fecundity in uniform age populations, accounting for a large amount of the variance in fecundity

observed. I also quantified the direct and indirect effects of body size on fecundity of wild Culex

nigripalpus using a path analysis. This analysis revealed that there is a strong indirect effect of

body size on fecundity mediated by bloodmeal size.

In experiments incorporating the effect of age, the responses changed somewhat. In Culex

nigripalpus, there was an interaction between age and bloodmeal size predicting fecundity,

indicating that the response to bloodmeal is not uniform across different aged groups, but rather

declines with age. Overall, the number of eggs produced by this species was found to decline

with age. This is the first demonstration of an age-related fecundity decline in Culex nigripalpus.

In Culex quinquefasciatus, the response of fecundity to body size is strong and positive in

younger age groups, declining as the mosquitoes age. Fecundity overall also declined with age

in this species.









These results demonstrate that fecundity declines with age, and increases with bloodmeal

size and body size. Responses to bloodmeal size and body size are modified by age. These

results can be incorporated into population growth models of these two species, which may aid in

better predicting risk of arboviral outbreaks.









CHAPTER 1
INTTRODUCTION

Mosquitoes

Taxonomy

The genus Culex is a sizable genus of the Culicidae, represented in North America by 29

species. Florida is home to 15 species in this genus. The two species considered in this thesis

are Culex nigripalpus Theobald and Culex quinquefasciatus Say. These fall in the subgenus

Culex Linnaeus (Darsie and Ward 2004).

The taxonomic status of Culex nigripalpus is uncontroversial, having been described by

Fred V. Theobald in 1901 (Knight and Pugh 1973), and undergoing no taxonomic revision since

then, except for addressing synonymy.

Culex quinquefasciatus on the other hand, has had quite a colorful history, and even today

its specific status is in doubt. At various times it has been considered a subspecies or geographic

variant of Culex pipiens Linnaeus, a species, or some combination thereof. For a thorough

treatment on the history of these taxa, see Vinogradova (Vinogradova 2000). For the purposes of

this paper, Culex quinquefasciatus will refer to Culex quinquefasciatus in the sense of Say 1823.

Distribution and Notes on Natural History

Culex nigripalpus (Fig. 1-1 A) is an abundant mosquito in the Southeastern USA and

ranges into Mexico, the Caribbean Basin and into Central and South America. Culex

quinquefasciatus (Fig 1-1 B) is also abundant in the Southeastern USA, but also ranges westward

into California. Outside the US, the range could be considered worldwide in warmer regions of

the globe. Both are warm weather species, and are found only in warm temperate, subtropical

and tropical climates. They are anautogenous species, requiring a bloodmeal to provision their

eggs.










Eggs are laid together on the surface of the water in batches referred to as rafts, which have

a characteristic structure (Figure 1-2). In general the egg rafts of both species are broadly oval,

range from 2-4mm in length, and are between 1-1.8mm in width, with 4-5 rows of eggs (Chadee

and Haegerl986). In general, all of the eggs matured in each gonotrophic cycle are laid in one

raft, although some may be retained and resorbed (Nayar and Knight 1981).

In terms of oviposition site selection, Culex quinquefasciatus prefers standing water bodies

high in organic pollutants, whereas Culex nigripalpus could best be described as a floodwater

mosquito, preferring freshly flooded grasslands and agricultural areas (Nayar 1982). Both

species will also oviposit in other types of habitats, especially containers. It is unknown what

drives habitat preferences in these species, although olfactory cues are likely to be important

(McCall and Eaton 2001, Olagbemiro et al. 2004).

Vector Associations

Mosquitoes of the genus Culex are well-known vectors of arboviral and other diseases in

humans and animals (Dow et al. 1964, Nayar 1982, van Riper III et al. 1986, Nayar et al. 1998,

Ahid et al. 2000, Turell et al. 2001). The first instance of incrimination of a Culex species as a

vector of an arboviral disease of humans was in 1933, during an outbreak of viral encephalitis in

the city of St. Louis, Missouri (Mitchell et al. 1980). Since that time Culex species have been

recognized as important vectors of many serious human ailments such as Japanese Encephalitis

in Asia (Rosen 1986), West Nile Virus (WNV) in Africa, Eurasia and the Americas (Sardelis et

al. 2001) and St. Louis Encephalitis (SLEV) in the Americas (Day 2001).

The species under consideration in this work are demonstrated vectors of SLEV and WNV.

Culex nigripalpus in particular is generally considered to be the most important vector for SLEV

and WNV in the State of Florida (Day 2001, 2005).









Understanding the role a vector plays in transmission cycles of an arbovirus or other

disease demands not only an understanding of vector competence, but also an understanding of

the reproductive rates of the vectors. Viral transmission intensity is related to size of the vector

population in many arboviral transmission systems (Mitchell et al. 1980, Day 2001), and thus

understanding factors that govern population size and increase is important to understanding the

viral cycles.

Age

Age of insect vectors of disease is generally investigated because of the relationship

between age of vector populations and the probability of disease transmission (Mitchell 1983).

Most vector-borne diseases require a period of time following exposure of the vector to reach a

stage where the vector is capable of transmission. This is termed the extrinsic incubation period

(Meyer 1989), and involves acquisition of the pathogen, propagation and development in the

vector, and spread of the pathogen into tissues that allow it to be transmitted (e.g., salivary

glands). The likelihood of transmission of a disease is then influenced by the proportion of the

vector population that may have been exposed to the disease and survived the extrinsic

incubation period. If the population is older, it is more likely that there are substantial numbers

of mosquitoes who have been exposed to the agent and survived through the extrinsic incubation

period.

Aside from vector potential of a population, age has other population-level and individual-

level effects. In many, if not most mosquito species investigated to date, age has been

demonstrated to have a negative impact on fecundity (Jalil 1974, Walter and Hacker 1974, Akoh

et al. 1992, Ferguson et al. 2003, Mahmood et al. 2004). These studies have investigated age

related effects on the fecundity of Culex quinquefasciatus, Culex tarsalis, Aedes triseriatus and

Anopheles stephensi.










The processes by which age negatively affects life history characteristics such as fecundity

are collectively referred to as senescence. Senescence is a phenomenon that is common to most

organisms (Kirkwood and Rose 1991), and is an umbrella term that encompasses many

physiological, genetic, and behavioral changes. The question as to why organisms undergo

senescence has been described in evolutionary terms (Gavrilov and Gavrilova 2002). A common

explanation is the antagonistic pleiotropy hypothesis. The antagonistic pleiotropy hypothesis

states that genetic mechanisms that enable high early-life fecundity or survival may also have the

result of lowering late life fecundity or survival. These effects would be especially pronounced

in organisms with a high level of per diem extrinsic mortality such as mosquitoes (Dow 1971,

Walter and Hacker 1974, Reisen et al. 1991). Thus, one could hypothesize pronounced effects of

aging on reproduction in mosquitoes.

An alternative to the antagonistic pleiotropy hypothesis is the mutation accumulation

hypothesis (Charlesworth 2000). This theory states that the gradual accumulation of the effects

of deleterious mutations over the lifespan of the individual is the reason we observe senescence.

Because these have pronounced effects only late in life, there is virtually no selection against

them. Note that this theory and the antagonistic pleiotropy hypothesis are not mutually

exclusive, and in fact predict the same high rate of senescence in short-lived creatures such as

mosquitoes.

Others have found that host seeking and oviposition activities of Culex nigripalpus are

inhibited during periods of low humidity and rainfall (Boike 1963, Provost 1969, Day et al.

1989, Day et al. 1990 a). During a drought, mosquitoes seek out sheltered locations to rest.

During this time, the age structure of the adult mosquito population may become old-biased.

Mortality and decreased fecundity in older population may prevent populations of these vectors









from recovering from such drought-induced quiescence (inactivity of mosquitoes during periods

of low rainfall and humidity) (Shaman et al. 2003). This would have an impact on sustained

arboviral transmission by reducing the abundance of vectors. If, after the amplification phase of

the transmission cycle there are too few vectors to continue transmission, enzootic maintenance

would be unlikely.

Bloodmeal Size

Several other physiological factors play a role in determining the production of eggs in

mosquitoes. Probably the most important factor in many species is the size of the bloodmeal

obtained. The relevance of this factor is easy to see when one considers that the bloodmeal is the

main source of nutrition a female mosquito uses to provision her eggs with yolk. The

relationship between bloodmeal size and fecundity is usually strong, and has been established in

various studies to follow a linear relationship (Miura and Takahashil1972, Edman and Lynn

1975, Akoh et al. 1992, Briegel 2003, Lima et al. 2003, Fernandes and Briegel 2005).

The linear relationship between bloodmeal size and fecundity has been questioned, as

undoubtedly there is an asymptote at the upper end of fecundity representing the maximum

number of ovarioles in the mosquito ovary (Miura and Takahashil1972). Another suggested

explanation for non-linearity in bloodmeal-fecundity regressions is that the non-linearity

represents a diminishing returns function of increasing bloodmeal mass on fecundity gain

(Roitberg and Gordon 2005). Roitberg and Gordon fit a constrained, quadratic function to their

data, but do not provide an estimate of the improvement in R2 ValUeS OVer a linear model.

Bloodmeal Source

Another factor that may be important in determining the possible fecundity of mosquitoes

is the source of the bloodmeal. Different hosts provide different nutritive values for feeding

mosquitoes, and this can determine the number of eggs matured and laid following a bloodmeal










(Mather and DeFoliart 1983). In Culex nigripalpus, Nayar and Sauerman (Nayar and Sauerman

1977) showed that egg production varied with host type, but no generalizations can be made.

The immune status of a host also may play a role in the expected fecundity return from a

given bloodmeal. It has been shown that fecundity from bloodmeal sources carrying parasites is

less than that from similar, uninfected sources (Ferguson et al. 2003, Lima et al. 2003). In

addition, antibodies to mosquito saliva may inhibit bloodmeal digestion by female mosquitoes

(Ramasamy et al. 1988), thereby reducing the expected fitness gain.

Body Size and Teneral Reserves

Body size of mosquitoes has been shown to be positively correlated with higher fecundity

(Miura and Takahashil972, Packer and Corbet 1989, Briegel 1990 b, 1990 a, Akoh et al. 1992,

Bradshaw and Holzapfel 1992, Blackmore and Lord 1994, Renshaw et al. 1994, Blackmore and

Lord 2000, Briegel and Timmermann 2001, Armbruster and Hutchinson 2002, Telang and Wells

2004). There are several reasons this may be the case. Larger body size is usually the result of

greater larval nutrition (Timmermann and Briegel 1993, Blackmore and Lord 2000, Briegel

2003), and hence the increase in fecundity with body size indicates mobilization of general

reserves for first-cycle reproduction. Another reason body size affects fecundity is that the

maximum number of ovarioles is generally greater in larger insects (Colless and Chellapah 1960,

Bonduriansky and Brooks 1999).This means that the larger the body size of an individual

mosquito, the more eggs she can mature given a bloodmeal of optimal size and quality. The

effect of body size on maximal number of ovarioles may only be seen in individuals who take

the maximum quantity of blood their midguts can allow.

In addition, larger mosquitoes can take larger bloodmeals, and this shows up in many

studies as a correlation between body size and bloodmeal size (Hogg et al. 1996, Lima et al.










2003, Fernandes and Briegel 2005). This relationship, when present, can vary in strength

(correlation coefficient), but is always of a positive slope.

Multiple Factors

Several studies have been done that investigate the effects of multiple factors on egg

production of mosquitoes. In general, multiple regression and MANOVA models improve

detection of effects and predictive capability because several factors are taken into consideration

at once. In the literature investigating the effects of age along with other factors that affect

fecundity, there has been no attempt to test for interactions between factors. It could be

hypothesized that the relationship of bloodmeal volume to fecundity, or body size to fecundity

might change as a result of age, reflecting physiological changes associated with senescence.

One example of these might be a change in the efficiency of bloodmeal digestion or yolk transfer

as mosquitoes age, although this has not been reported. Another mechanism may be the

depletion of general protein reserves. Insects store protein derived from larval nutrition as

hexamerins (Telang et al. 2002), and these stores are generally greater in larger, better fed

insects. If storage protein decreases non-linearly in relation to size with age, it could show up as

an interaction between body size and age predicting fecundity.

Summary

The series of experiments reported in this thesis were designed to determine the

contributions of body size, bloodmeal size and age to the fecundity of Culex nigripalpus and

Culex quinquefasciatus. The results of these experiments are of interest from a practical

standpoint because the estimates of fecundity reduction with age may be incorporated into

models of the population dynamics of these species, which might improve predictions of

outbreaks of arboviral diseases.










They are also of interest from the standpoint of improving knowledge about the ecological

and physiological parameters governing reproductive success.
































A. Culex nigripalpus Theobald.


Figure 1-1.


B. Culex quinquefa~sciatus Say.










































Figure 1-2. Egg raft of Culex quinquefasciatus.
darkened.


ese eggs are freshly laid, and have not









CHAPTER 2
EFFECTS OF BLOODIVEAL SIZE AND BODY SIZE ON FECUNDITY OF WILD Culex
nigripalpus

Introduction

Culex nigripalpus Theobald is a widespread nuisance mosquito found throughout the

Neotropics. This mosquito is also known to breed in large numbers throughout Florida and parts

of the Southeastern U.S. It is a capable vector of West Nile Virus (Sardelis et al. 2001) and St.

Louis Encephalitis Virus (Day 2001); both arboviruses are in the family Flaviviridae (Calisher

and Karabatsos 1988).

Due to its status as a vector of arboviruses, this species of mosquito has been studied

intensively for many years (Nayar 1982). Factors governing its seasonal population cycles have

been of great interest, as vector abundance during critical periods of the year can be essential to

the continuous transmission of pathogens (Day 2001). Reproductive output in mosquitoes is

known to be dependent on several factors such as body size, bloodmeal size, bloodmeal source,

infection status and age (Edman and Lynn 1975, Akoh et al. 1992, Lima et al. 2003).

Given that resources allocated to reproduction in anautogenous mosquitoes come from the

bloodmeal, larval diet and to some degree sugar feeding in the adult stage (Nayar and Sauerman

1975, Timmermann and Briegel 1999, Briegel 2003), one would predict that these factors would

be important predictors of fecundity. This has in fact been demonstrated in many experiments.

Few studies have investigated the effect of these factors in wild-caught mosquitoes. In addition,

many of these studies have shown that there is some dependence between body size and

bloodmeal size, yet in the creation of predictive models this dependence has not been addressed.

If body size is responsible for an increased intake of blood, and this translates into increased

fecundity, then larval nutrition could be regarded as having a greater contribution to lifetime

fitness than previously imagined. Data are needed to better quantify the resource inputs










important to the maintenance and growth of vector populations, and incorporate these into

models of the population dynamics of Culex nigripalpus.

This study was designed to demonstrate the relationship between body size, bloodmeal size

and fecundity, and also to separate the effects of body size into direct effects, representing the

influence of larval nutrition, and indirect effects, representing the influence of body size on

bloodmeal size.

Materials and Methods

Trapping

Host seeking females were collected in the field using two lard can traps hung 1.5 m off

the ground at Lockwood Hammock near Vero Beach (27.575720N, 80.436180W). This site is

well known for producing large numbers of Culex nigripalpus. The traps were baited with a live

chicken (Production Red strain) placed within the trap in a mesh bag, allowing trapped

mosquitoes to bloodfeed (University of Florida IUCUC # D509). The traps were set from 6pm

to 8am on June 1, 2006. This timeframe takes account of the evening and morning host-seeking

and bloodfeeding habits of this species.

Hematin Collection and Analysis

Bloodmeal size can be quantified by measuring the amount of hematin in the excreta

(Briegel 1980). Hematophagous arthropods void acid hematin as a byproduct of hemoglobin

digestion. It has been found that the quantity of hematin voided correlates in a linear manner

with the amount of blood ingested (Briegel 1980, 2003). Therefore, it is an appropriate means of

quantifying the relative amount of blood ingested. Other methods including near infrared

spectrometry (Hall et al. 1990) and weighing of mosquitoes before and after a bloodmeal

(Roitberg and Gordon 2005) were rej ected as being too time-consuming, invasive, or requiring

equipment that was unavailable.









Following trapping, the mosquitoes were brought to the laboratory, and 108 bloodfed

females were selected and placed into separate 40 ml vials covered with screen. 10% sucrose

was provided on small cotton balls, and the females were given four days to mature eggs in the

vial. Temperature was held constant at 27.60 C and relative humidity was 70%.

The time required for digestion and egg development by this species has been found to be

72 h at 30oC, and 96 h at 24oC (Nayar and Knight 1981), and hence four days at 27 oC was found

to be a good balance between maximal survival in the vials and maximal ovarian development.

Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00 ml

1% LiCO3. The resulting solutions were decanted into spectrophotometric cuvettes and the

absorbance at 387 nm was read in a spectrophotometer. Absorbance readings were converted to

micrograms of hematin using a standard curve previously prepared (Appendix B).

Oviposition

After egg maturation, gravid females were transferred into a second set of 40ml vials

containing 4.0 ml of 10% (by volume) hay infusion in tap water for oviposition. These vials

were placed in a screened outdoor enclosure at approximately 5pm. This was done to provide a

natural twilight which has been found to induce greater oviposition rates in this species.

Oviposition continued for three nights. Each morning, egg rafts deposited were removed, placed

on water under a microscope and photographed at high magnification with a digital camera.

Photographs of the egg rafts were printed out with a standard laserj et printer and the number of

eggs counted (method suggested by A. Doumboya, personal communication, 2005).

Winglengths

Following oviposition, females were removed, killed, identified to species (Darsie 2004),

and their wings removed for measurement. Abdomens were dissected to count retained eggs.

The wings were measured by photographing them adj acent to a steel pin of known length









(measured to the nearest thousandth of a millimeter). The photographs were opened in

SigmaScan (SPSS Inc.), scaled using 2-point rescaling, and measured from the alular notch to

the distal end of R2, excluding fringe hairs. (Figure 2-1).

Analyses

All statistical analyses were conducted using S-Plus@ 7.0 for Windows@ (Insightful

Corp.). Fecundity was scored as the number of eggs in the egg raft plus the number of retained

eggs, provided the retained eggs numbered fewer than 50. If females retained greater than 50

eggs, they were discarded in the analysis, due to difficulty in counting large numbers of retained

eggs (eggs often burst, obscuring the slide with opaque yolk).

The effect of bloodmeal size (measured by hematin) and body size (measured by

winglength) on fecundity was analyzed using simple linear regressions and a multiple regression

of these two factors plus their interaction on fecundity. Non-significant effects and interactions

were discarded in a stepwise fashion.

Two additional regressions were performed in which the centered, standardized scores (Z-

scores) of bloodmeal size and body size were used as the predictors (Marquardt 1980), and the

unstandardized fecundity as well as the standardized fecundity were used as the responses.

To explore the notion that body size has both direct and indirect effects on fecundity, a

path analysis was conducted on the fully standardized multiple regression. The hypothesis was

that the correlation between body size and bloodmeal size implies that body size has a direct

effect on fecundity, and an indirect effect mediated by bloodmeal size. In this path analysis,

body size was considered an exogenous variable and bloodmeal size an endogenous variable.

The path analysis presents a causal hypothesis about the relationship between the variables, and









then quantifies the effects in the hypothesis. It does not imply that the causal hypothesis is

correct (Darlington 1990).

Assumptions of homoscedasticity and normality were checked graphically with plots of

residuals versus fits and Q-Q plots of the residuals respectively. A one sample Kolmogorov-

Smirnov Test of Composite Normality was also performed to verify normality of the residuals.

Results

Only one of the Culex that oviposited in the vials was not Culex nigripalpus. This was a

female Culex quinquefasciatus, and was not used in the analyses.

In total, 84 female Culex nigripalpus survived and oviposited in the vials. Of these, 23

retained at least one egg. Overall, egg retention was low, averaging just 1.8 eggs/female. Two

females retained more than 50 eggs, and these were not included in the analyses. A summary of

data collected on fecundity, winglength and hematin excreted can be found in Table 2-1.

Both factors (body size and bloodmeal size) were significant in all regression analyses at

the a=0.05 level. Regression equations can be found in Table 2-2. Scatterplots of the

unstandardized simple linear regressions are shown in Figures 2-2 and 2-3.

What is evident from the standardized regression equation is that one standard deviation in

winglength is responsible for a greater gain in fecundity than a standard deviation in bloodmeal

size. With the unstandardized regression, such an interpretation is not intuitive. The

unstandardized regression is given because the equation can be used to predict fecundity by

inputting normal measures of body size and bloodmeal size.

Path Analysis

The path analysis indicates that the total contribution of body size to fecundity can be

decomposed into a direct effect, and an indirect effect mediated by bloodmeal size (Table 2-4,

Fig. 2-4). This is a quantification of a causal hypothesis, not a proof of such a hypothesis.









Discussion

The mean fecundity for female Culex nigripalpus in this study was quite high (273)

compared to the values obtained in 1975 from Tiger Hammock near Vero Beach(Edman and

Lynn 1975) (175) and 1977 (Nayar and Sauerman 1977) (160). There are several possibilities

for why this difference in fecundity from previous studies was noted:

Greater size of bloodmeal: The mosquitoes in this study had access to a host for over 14
h, whereas in the studies cited, the time allowed for feeding was not stated.
Strain differences: Perhaps the innate fecundity of the mosquitoes captured in this study
was higher than the previous ones cited
Strain differences in the chicken: Perhaps the blood of the chicken used in this study had
greater protein content. Unfortunately the strain used in the previous studies was not
recorded.
Previous bloodfeeding: These mosquitoes, being wild caught, may have fed previously
and thus had more usable protein to mature eggs than ones raised in the laboratory.
This is another instance of bloodmeal size being a strong predictor of fecundity in a

mosquito. This is easily understood, as the bloodmeal is the main source of protein for the

development of mature oocytes (Briegel 2003). Edman and Lynn showed that bloodmeal

volume is positively correlated with egg maturation in Culex nigripalpus (Edman and Lynn

1975), and many others have shown the same trend in other species of anautogenous mosquitoes

(Colless and Chellapah 1960, Cochrane 1972, Miura and Takahashil972, Hurd et al. 1995, Lima

et al. 2003, Fernandes and Briegel 2005, Roitberg and Gordon 2005). This regression simply

provides a more precise accounting of the relationship between bloodmeal size and fecundity in

this species.

The multiple regression predicting fecundity from bloodmeal size and body size provides a

better estimate of potential fecundity than does either factor alone. This reflects the fact that

there are two sources of protein for the development of at least the first batch of eggs in any

autogenous mosquito: larvally-acquired protein, and protein from the bloodmeal. Taking both









into consideration provides a more precise estimate of the net contribution any given female is

likely to make to a particular egg batch.

The path analysis of the multiple regression indicates a direct contribution of body size and

an indirect effect mediated by bloodmeal size. This would indicate that body size contributes to

fecundity, probably due to the contribution of general reserves. It also contributes by enabling a

greater volume of blood, and hence protein, to be ingested during bloodfeeding. This suggests

that one of the most important factors determining a mosquito' s fecundity is the quality of the

larval environment. Higher-quality environments (meaning more nutrients available) will result

in larger adults with greater reserves (Telang and Wells 2004), able to take larger bloodmeals,

and producing greater numbers of eggs. This type of relationship should be evident in the wild as

Culex nigripalpus varies in size over the year due to changes in development time brought about

by water temperature changes (Day et al. 1990 b).

Another way that body size might have a positive influence on fecundity is by increasing

the maximum number of ovarioles that can be used to produce eggs. It is known that the number

of ovarioles in insects is influenced by size, and so larger individuals have a higher maximum

number of eggs that can be matured each gonotrophic cycle (Fitt 1990). This would be evident

in mosquitoes taking a bloodmeal of maximum size and quality, where the great maj ority of

follicles are provisioned with yolk.

The total fitness effects of larval habitat quality would be greater than those analyzed in

this experiment, as it would affect larval survival (Agnew et al. 2000, Reiskind et al. 2004).

There is also evidence that larger adult mosquitoes live longer (Nasci 1986, Lounibos et al. 1990,

Suzuki et al. 1993), and thus size would contribute substantially more to overall fitness than this

analysis has shown.

























Figure 2-1. Example of a wing photograph used to measure winglengths of female Culex
nigripalpus. Arrow indicates length of measurement, from alular notch (Al) to distal
end of wing vein R2-


Table 2-1. Summary statistics for bloodmeal size, body size and fecundity of wild Culex
nigripalpus used in analyses.
Hematin excreted Winglength Fecundity Retained
(Clg) (mm) eggs
Minimum 7.21 2.41 148 0
Mean 18.30 2.95 273 1.83
Maximum 34.57 3.46 385 40
Std. Dev 5.87 0.22 57.13 6.23










Table 2-2. Summary of four regressions testing relationships between hematin excreted,
winglength and fecundity. Significance values are for the t-statistic. F-tests for all
models were below 0.05.


Model


Parameter Estimate Std.
Error
Intercept 144.93 14.35
Hematin 7.00 0.75


DFt T
stati stic
1,82 10.10
9.37

1,82 -5.63
1061

2,81 -4.01
5.78
7.07

2,81 79.21
5.78
7.07



2,81 0
5.78
7.07


Hem atin


<0.01
<0.01

<0.01
<0.01

<0.01
<0.01
<0.01

<0.01
<0.01
<0.01


1.00
<0.01
<0.01


0.52


0.58


0.70


Winglength


Stepwise model*,
unstandardized



Stepwise model*,
regressors
standardized, response
unstandardized

Stepwise model*, all
parameters
standardized


Intercept
Winglength

Intercept
Hematin
Winglength

Intercept
Hematin
Winglength


Intercept
Hematin
Winglength


-310.85
197.71

-202.25
4.14
135.29

273.07
24.29
29.74


0
0.4251
0.5205


55.19
18.63

50.37
0.72
19.12

3.44
4.20
4.20


0.603
0.0736
0.0736


0.70


0.70


*The interaction between the predictors was not significant at the a
removed as part of a stepwise model reduction.
? Indicates degrees of freedom for overall model F-test


=0.05 level, and hence was











































** *


400


C.
--350


-
300 -







1 50



100 -


* *


*81




** Fecundity=144.93 + 7.00*Hematin


30 35


Figure 2-2. Scatterplot of fecundity versus unstandardized hematin for wild Culex nigripalpus.
The line represents the least-squares linear regression.


15 20 25

Hem~atin Excreted (gLg)





3 -
11Il l

2 0


Fecundity---310.85 + 197.71*Winglength *


I
Ir
r
=~


~


5; *
rr


3 2.5 2 7


3 3 3.5


Figure 2-3. Scatterplot of fecundity versus unstandardized winglength for wild Culex
nigripalpus. The line represents the least-squares linear regression.


2 9 3 1

Winglength (mm)




































Table 2-3. Decomposition of effects in path analysis. Note that the direct effect of bloodmeal
size is the portion of its contribution not attributable to body size.
Factor Direct Indirect Spurious Total Pearson
effect effect effect effect Correlation
coefficient
Effects on Body size 0.520 0.240 0 0.761 0.761
Fecundity Bloodmeal 0.425 0 0.294 0.425 0.719
Size
Effect on Body Size 0.565 0 0 0.565 0.565
Bloodmeal


0- 2049




Fecundity


Body Size












Bloodmeal
Size


Figure 2-4. Path diagram describing direct effects between the bloodmeal size, body size and
fecundity .









CHAPTER 3
INFLUENCE OF BLOODMEAL SIZE AND BODY SIZE ON THE FECUNDITY OF
C AP TIVE Culex quinque fasciatus

Introduction

Culex quinquefasciatus is a ubiquitous peridomestic mosquito in urban and rural

environments across the tropics and subtropics. It is a known vector of various filariases

(Villavaso and Steelman 1970, Lowrie et al. 1989) as well as viral diseases (Mitchell et al. 1980,

Sardelis et al. 2001) and avian malarias (van Riper III et al. 1986). Because of its importance in

the transmission of pathogens, considerable interest has been shown in the bionomics of this

species.

One measure of the fitness of individual animals is their fecundity, or reproductive output.

Because Culex mosquitoes lay eggs in rafts, usually depositing their entire clutch at once, one

can easily estimate the reproductive output of a mosquito by counting the number of eggs in the

egg raft. This is the measure is most commonly used, and relies on the assumption that most

mosquitoes in the wild deposit only one clutch (due to high adult mortality).

Factors known to affect the fecundity of this species include bloodmeal size and body size

(Lima et al. 2003). Bloodmeals provide the protein as well as some of the carbohydrate and lipid

needed to provision a clutch of eggs (Clements 2000 a). In anautogenous mosquitoes, bloodmeal

size is almost always correlated positively with fecundity (Cochrane 1972, Miura and

Takahashil1972, Edman and Lynn 1975, Mahmood et al. 2004).

From a theoretical standpoint, body size may correlate with fecundity via several causal

mechanisms. The first is that the protein reserves of larger mosquitoes are generally higher

(Briegel 2003). The presence of higher general protein reserves means that more protein may be

diverted to egg development regardless of the size of the bloodmeal. The second way that body

size may be related to fecundity is that larger insects generally have larger numbers of ovarioles










(Fitt 1990), thus the maximum number of eggs that can be matured given an optimum bloodmeal

is greater. In any case, the size of adult female mosquitoes is thought to be principally the result

of greater food resources available during larval development. There is also evidence that

survival and reproductive success is influenced by size of adult female mosquitoes, so the overall

contribution of body size to lifetime fitness may be considerable (Nasci 1986, Lounibos et al.

1990, Suzuki et al. 1993).

Experiments have in the past considered bloodmeal size and body size separately, not

addressing their combined contribution to fecundity of the mosquitoes. The experiment

described here was designed to determine the combined effect of these two factors on fecundity

of Culex quinquefasciatus using multiple regression analysis.

Materials and Methods

Larval Rearing

Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing

approximately 700 ml of tapwater. Pans were set with 3 egg rafts each of colonized Culex

quinquefasciatus, from USDA ARS Gainesville FL, established 1995 (Allan et al. 2006). Food

was provided daily to each pan as 20 ml of slurry containing 20 mg/ml 1:1 Brewer' s yeast/liver

powder. This rearing regimen was chosen in an attempt to generate a range of mosquito sizes,

while still achieving relatively simultaneous emergence as adults.

Pupation and Bloodfeeding

Pupae were placed in 500 ml cups containing 100 ml of tapwater, and allowed to emerge in

a cage measuring 33 x 33 x 33 cm. Adults were provided with 10% sucrose solution on a cotton

wick replaced daily. After 7 days, the mosquitoes were offered a bloodmeal on a restrained

chicken (University of Florida IUCUC # D509). Mosquitoes were allowed to feed to repletion.









Hematin Collection and Quantification

Bloodmeal size can be quantified by measuring the amount of hematin in the excreta.

Hematophagous arthropods void acid hematin as a byproduct of hemoglobin digestion. It has

been found that the quantity of hematin voided corresponds in a linear manner with the amount

of blood ingested (Briegel 1980, 2003). Therefore, it is an appropriate means of quantifying the

relative amount of blood ingested. Other methods including near infrared spectrometry (Hall et

al. 1990) and weighing of mosquitoes before and after a bloodmeal (Roitberg and Gordon

2005)were rej ected for this study as being too time-consuming, invasive, or requiring equipment

that was unavailable.

Immediately following bloodfeeding, individual mosquitoes placed in separate 40 ml vials

2.5 cm diameter, 9.5 cm deep) covered with screen. 10% sucrose was provided on small cotton

balls. Following a 4 day period for egg maturation, females were transferred to separate vials

(see below) for oviposition. The time required digestion and egg development for this species

has been found to be 2-3 days (Elizondo-Quiroga et al. 2006), and hence four days was found to

be a good balance between maximal survival in the vials and maximal ovarian development.

Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00 ml

1% lithium carbonate. The resulting solutions were decanted into spectrophotometric cuvettes

and the absorbance at 387 nm was read in a spectrophotometer. Absorbance readings were

converted to micrograms of hematin using a standard curve previously prepared (Appendix B).

Oviposition

After egg maturation, gravid females were transferred into another set of 40ml vials

containing 4.0 ml of 10% hay infusion (by volume) in tap water for oviposition. The following

morning, egg rafts deposited were removed, placed on water under a microscope and

photographed at high magnification with a digital camera. Photographs of the egg rafts were










printed out using a standard laserj et printer and the number of eggs counted (method suggested

by A. Doumboya, personal communication, 2005).

Egg rafts were replaced in the vials and incubated at 27. 1C for 36 h and hatched larvae

were filtered onto white filter paper and counted. Percentage hatch was calculated as the number

of larvae hatched divided by the number of eggs multiplied by 100.

Winglengths

Following oviposition, females were removed, killed, and their wings excised and mounted

on slides for measurement. The wings on the microscope slides were photographed with a

standard size reference (a length of steel measured with a caliper to the nearest thousandth of a

millimeter). The photographs were opened in SigmaScan Pro 5 (Systat Software, Inc., Point

Richmond, CA), calibrated for size, and measured from the alular notch to the distal end of R2,

excluding fringe hairs (Packer and Corbet 1989). It was decided to use the distal end ofR2 as a

measurement point as it is an unambiguous standard landmark. Other studies have used the

distal end of the wing, or some kind of other subj ective measure of the maximal distance (Packer

and Corbet 1989, Lima et al. 2003). While others have suggested transforming the winglengths

thus obtained by cubing the linear measure (Briegel 1990 a), the recommendations of Siegel

were followed here (Siegel et al. 1992), and winglengths were not transformed.

Analyses

All statistical analyses were conducted using S-Plus@ 7.0 for Windows@ (Insightful

Corp.).

Summary statistics were calculated for each measured parameter, including minimum,

mean, maximum and standard deviation.

Fecundity was scored as the number of eggs in the egg raft. The effect of bloodmeal size

(measured by hematin) and body size (measured by winglength) on fecundity was analyzed










using simple linear regressions and a multiple regression of these two factors plus their

interaction on fecundity. If the interaction was not significant, it was discarded in creation of a

predictive multiple regression model. A multiple regression of fecundity on the standardized

values (Z-scores) of the predictors was also performed. Z-scores were calculated by subtracting

the mean value of a regressor from each observation, then dividing this by the standard deviation

of the regressor. This produces mean values of zero and standard deviations of one. When used

in multiple regression analysis, it allows a more standard interpretation of slope values, i.e. It

apportions mean changes in response due to predictor variations of one standard deviation

(Marquardt 1980). Doing so allows one to order the predictors in terms of influence on a

common scale.

Assumptions of homoschedasticity and normality were checked graphically with plots of

residuals versus fits and Q-Q plots of the residuals respectively. A one sample Kolmogorov-

Smirnov Test of Composite Normality was also performed to verify normality of the residuals.

A regression analysis testing for linear dependence between the predictors in the multiple

regression model was also performed.

Results

Winglength varied between 2.89 and 3.33mm with a mean of 3.10 and a standard deviation

of 0.10 mm (Table 3-1). Hematin voided was also variable, with a mean of 15.99Clg and a

standard deviation of 4.96 lg.

Significant simple linear regressions were found for each of the variables (Table 3-2). The

predictive power of hematin as a factor predicting fecundity was greater than that for the simple

linear regression of winglength predicting fecundity (Table 3-2).

The full multiple regression model predicting fecundity from both bloodmeal size and

body size had a much better fit to the data than either factor considered alone. The multiple










regression with the standardized variables as predictors indicated that for one standard deviation

in either variable, the effect of bloodmeal was the more significant source of variation in

fecundity .

A regression of hematin and winglength on percentage hatch was considered, but rej ected

due to non-normality of the data and no apparent trend apparent in scatterplots of the data (data

not shown).

Discussion

The amount of hematin voided by this group of mosquitoes was similar to that in a study

conducted on a wild strain from Brazil (Lima et al. 2003). In that study the mean hematin

content in excreta varied from 14.60-15.80 Clg (fed on human blood). Mean fecundity for the

Brazilian strain was also much lower, perhaps indicating that the usable protein content for a

given quantity of hemoglobin is less in human blood than in chicken blood. A study of Culex

nigripalpus showed that mean number of eggs per raft was greater with chicken blood than with

human blood (Nayar and Sauerman 1977). Alternatively, this could indicate an innate difference

in the fecundity of the two strains of mosquito.

The slope of the regression of winglength on fecundity was about twice as great for this

strain than for a Brazilian strain fed on human blood, although the mean winglength of the

Brazilian strain was substantially greater than this population (Lima et al. 2003). This may be

partially the result of a slightly different winglength measurement technique.

The combined effect of larval nutrition (represented by adult size) and bloodmeal size

(represented by hematin) on fecundity was greater than either of the two factors alone. In this

case, bloodmeal was seen to be a better predictor of fecundity than adult body size, but this may

be due to the fact that the mosquitoes raised did not vary greatly in body size. A comparison

with known body sizes of wild Culex quinquefasciatus was not possible. Other studies indicate










that the adult body size in this species is larger (Lima et al. 2003), but these may represent strain

differences.

It seems likely that the maj or contributors to female fecundity of this species are

bloodmeal volume, bloodmeal source and general reserves. This experiment demonstrates that

excluding sources of mortality, fitness of Culex quinquefasciatus is predicated upon bloodmeal

size and general size.










Table 3-1. Summary statistics for data measured in regression analyses.
Parameter
Statistic Hematin (Clg) Winglength (mm) Fecundity (# of percentage
eggs) hatch
Min 5.57 2.89 76 7
Mean 15.99 3.10 163 85
Max 30.51 3.33 256 100
N 71 71 71 71
Std. Deviation 4.96 0.10 36 20

Table 3-2. Summary of four regressions testing relationships between hematin excreted,
winglength and fecundity. Significance values are for the t-statistic. F-tests for all
models were below 0.05.


Model


Parameter Coefficient

Intercept 79.79
Hematin 5.20


Std.
Error
10.31
0.62


DFT T
stati stic
7.74
1,69
8.44


P RL


<0.01
<0.01

0.10
<0.01

<0.01
<0.01
<0.01


Hem atin


0.51


0.11



0.65


Intercept
Winglength

Intercept
Hem atin
Winglength


-200.19
117.10

-319.35
5.319
128.10


Winglength


121.19 -1.65
1,69
39.06 3.00


78.16
0.53
24.92


-4.09
2,68 10.09
5.14


Stepwise model*,
unstandardized


Stepwise model*, Intercept 162.93 2.59 2,68
regressors standardized, Hematin 26.36 2.61
response unstandardized Winglength 13.43 2.61
*The interaction between the predictors was not significant at the a=0.05
removed as part of a stepwise model reduction.
? Indicates degrees of freedom for overall model F-test.


62.86 <0.01 0.65
10.09 <0.01
5.14 <0.01
level, and hence was





,*



*


*


200 -








- 15


100





50


/*


**


Figure 3-1.


Scatterplot of fecundity versus bloodmeal size for Culex quinquefasciatus. Data are
presented in unstandardized format. Least squares regression line shown.


15 20

Hematin (p~g)

















250









S150 *




100 *





2.8 2.9 3.0 3.1 3.2 3.3

Winglength (mm)


Figure 3-2. Scatterplot of fecundity versus winglength for Culex quinquefasciatus. Data are
presented in unstandardized format. Least squares regression line shown.









CHAPTER 4
Culex nigripalpus AGING AND FECUNDITY

Introduction

Culex nigripalpus is a species of mosquito distributed throughout the Neotropics and in

large portions of the Caribbean Basin and the Southeastern United States (Nayar 1982). It has

been incriminated as a vector of St. Louis Encephalitis Virus (Dow et al. 1964) and West Nile

Virus (Sardelis et al. 2001), as well as several other viral, protozoan, and helminth diseases of

man and animals (Nayar 1982, Nayar et al. 1998). It is believed that this species is the major

enzootic vector of both WNV and SLE in Florida (Day 2001, Rutledge et al. 2003). Considering

its role in the transmission of arboviruses in Florida and other places, much attention has been

focused on the bionomics and population dynamics of this species (Dow 1971, Nayar 1982, Day

et al. 1990 b, Day 2001).

Factors affecting the fecundity (egg production) of mosquitoes have been investigated for

many years. In general, larval nutrition, adult nutrition and age are considered important factors

(Clements 2000 b, Briegel 2003). A portion of the protein reserves acquired during larval

development may be used for the provisioning of eggs (Briegel 2003). Larval nutrition is

difficult to quantify in the field, but it has been shown that the size of an adult mosquito

correlates well with the amount of nutrients available to the larvae (Akoh et al. 1992, Blackmore

and Lord 1994, Agnew et al. 2000, Blackmore and Lord 2000, Armbruster and Hutchinson

2002). Because of this, adult size serves as a convenient proxy for general protein reserves.

In addition to general protein reserves, anautogenous mosquitoes also use the bloodmeal as

a source for protein used to mature eggs. Bloodmeal size has usually been highly correlated with

fecundity in many mosquito species studied to date, and this holds true for Culex nigripalpus










(Edman and Lynn 1975, Ferguson et al. 2003). Measuring both bloodmeal size and body size has

the potential to address the contributions of two important sources of variation on fecundity.

This effect of age on fecundity can be seen as an aspect of senescence, the umbrella term

encompassing all deleterious effects of time on fitness of an individual. Age has been

demonstrated to have an effect on fecundity in Culex mosquitoes (Walter and Hacker 1974,

Suleman and Reisen 1979, Akoh et al. 1992), but whether age modifies the response to

bloodmeal size or body size has not been investigated. It is conceivable that physiological

systems mediating fecundity such as digestion or mobilization of general reserves are impacted

by senescence.

In any case, generalizing the responses of species studied so far, fecundity should increase

with bloodmeal size and body size, and decrease with age of the mosquito. The purpose of this

experiment was to analyze the response of fecundity to these three factors (bloodmeal size, body

size, and age) combined, as well as any interactions in a linear multiple regression model.

Materials and Methods

Larval Rearing

Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing

approximately 700 ml of tapwater. Pans were set with 3 egg rafts each from a colony of Culex

nigripalpus. The colony was established from a Vero Beach FL collection (Allan et al. 2006) in

1999 (Erin Vrzal, USDA, personal communication). Food was provided daily to each pan as 20

ml of slurry containing 20 mg/ml 1:1 Brewer' s yeast/liver powder.

Pupation

Pupae were placed in 500 ml cups in a large cage measuring 57 x 57 x 57cm and sugar was

provided to the emerging adults as 10% sucrose solution on cotton wicks, replaced daily. Adults

were allowed to emerge for 12 h following the emergence of the first female, whereupon the









cups were removed. The final density of mosquitoes in the cage was estimated to be about 500

females. Immediately following the removal of the cups, the cage was placed in a separate room

with windows, as it has been noted that successful mating in this colony is inhibited if

mosquitoes are denied access to either sunlight or twilight (personal observation). Following 2

days in this separate room, the cage was returned to the controlled-environment chamber where

temperature was maintained at 26.0fl.90C for the duration of the experiment. Humidity was

88.5f 7.5% RH. Light cycle was 16:8 L:D.

Bloodfeeding

Beginning at 5 days post-eclosion, varying numbers of host-seeking females were removed

from the cage and bloodfed on a restrained chicken (University of Florida IUCUC # D509). This

was repeated at four day intervals until 6 groups were obtained. The oldest group was bloodfed

at 25 days post-eclosion. The numbers removed for bloodfeeding were adjusted upwards with

age to anticipate higher mortality of older individuals. Mortality was severe in the oldest age

group, with only 1 out of 66 bloodfed females surviving to oviposit, compared with 30 of 31

females surviving to oviposit in the youngest age group.

Termination of Study

The study was terminated when the combination of mortality and removal of host-

seeking females depleted the supply in the large cage. As a result, ages from 5-25 days post-

eclosion were tested. Age at oviposition was calculated by simply adding four days to the age at

bloodfeeding.

Hematin Collection and Quantification

Bloodmeal size was quantified by measuring the amount of hematin in the excreta.

Hematophagous arthropods void acid hematin as a byproduct of hemoglobin digestion. It has

been reported that the quantity of hematin voided by many mosquitoes corresponds in a linear









manner with the amount of blood ingested (Briegel 1980, 2003). Therefore, hematin analysis is

an appropriate means of quantifying the relative amount of blood ingested. Other methods

including near infrared spectrometry (Hall et al. 1990) or weighing of mosquitoes before and

after a bloodmeal (Roitberg and Gordon 2005) were rej ected for this study as being too time-

consuming, invasive, or requiring equipment that was unavailable.

Immediately following bloodfeeding, individual mosquitoes were placed in separate 40 ml vials

2.5 cm diameter, 9.5 cm deep) covered with screen. 10% sucrose solution was provided on small

cotton balls, replaced daily. Following a four day period for egg maturation, females were

transferred to separate vials (see below) for oviposition. The time required for digestion and egg

development for this species has been found to be 72 h at 30oC, and 96 h at 24oC (Provost 1969,

Nayar and Knight 1981), and hence four days at 27 oC was found to be a good balance between

maximal survival in the vials and maximal ovarian development.

Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00

ml 1% LiCO3. The resulting solutions were decanted into spectrophotometric cuvettes and the

absorbance at 387 nm was read using a spectrophotometer. Absorbance readings were converted

to micrograms of hematin using a standard curve previously prepared (Appendix B).

Oviposition

After egg maturation, gravid females were transferred into a second set of 40ml vials

containing 4.0 ml of 10% (by volume) hay infusion in tap water for oviposition. These vials

were placed in a screened outdoor enclosure at approximately 5pm. This was done to provide a

natural twilight which has been found to induce greater oviposition rates in this species (personal

observation). If females did not oviposit on the first night, they were allowed a second night to

oviposit. The following morning, egg rafts deposited were removed, placed on water under a

microscope and photographed at high magnification with a digital camera. Photographs of the









egg rafts were printed out using a standard laserj et printer and the number of eggs counted

(method suggested by A. Doumboya, personal communication, 2005).

Egg rafts were replaced in the vials in which they were laid and incubated at 27. 1oC for 36

h to allow hatching (Provost 1969). Hatched larvae were filtered onto white filter paper and

counted. Percentage hatch was calculated as the number of larvae hatched divided by the

number of eggs multiplied by 100.

Winglengths

Wings were photographed with a digital camera adj acent to a steel pin of known size. The

wing photographs were opened in SigmaScan Pro 5 (Systat Software, Inc., Point Richmond,

CA), calibrated for size using a 2-point rescaling function, and measured from the alular notch to

the distal end of R2, excluding fringe hairs (Packer and Corbet 1989). It was decided to use the

distal end of R2 aS a measurement point rather than the wingtip, as it is an unambiguous standard

landmark. While others have suggested transforming the winglengths thus obtained by cubing

the linear measure (Briegel 1990 a), the recommendations of Siegel were followed here (Siegel

et al. 1992), thus winglengths were not transformed.

Analyses

All regression analyses were conducted using S-PlusO 7.0 for Windowso (Insightful

Corp.). ANCOVA slope estimates were tested in SAS version 9.00 for Windows. Fecundity

was scored as the number of eggs in the egg raft plus the number of retained eggs in the killed

females, provided these numbered less than 50. If females retained more than 50 eggs, they were

discarded in the analysis, due to difficulty in counting large numbers of eggs (eggs often burst,

obscuring the slide with opaque yolk).

Simple linear regressions of each predictor on fecundity were performed to compare with

published results of these predictors on fecundity.










A multiple linear regression was performed on standardized variables regressing fecundity

on body size, bloodmeal size and age (Marquardt 1980). Non-significant interactions and main

effects (if appropriate) were discarded in a stepwise fashion. Z-scores were calculated by

subtracting the mean value of a regressor from each observation, then dividing this by the

standard deviation of the regressor. This produces mean values of zero and standard deviations

of one. When used in multiple regression analysis, it allows a more standard interpretation of

slope values (Marquardt 1980), i.e. It apportions mean changes in response due to predictor

variations of one standard deviation. It also allows for more adequate assessment of model

quality and provides better insight for stepwise model reduction. For comparative purposes, a

multiple regression of fecundity on the untransformed variables was also performed.

Normality was confirmed with a Kolmogorov-Smirnov Test of Composite Normality of

the residuals, and homoschedasticity was verified visually with a plot of residuals versus fitted

values.

Significant interactions with age were explored by examining scatterplots of the interacting

variables, and generating slope estimates by ANCOVA. The ANCOVA was set up using the

recommendations of Huitema (Huitema 1980), and was done to illustrate the differences in slope

of the simple regression of one variable predicting fecundity at different levels of the interacting

variable. The mosquitoes aged 21 days and 25 days were grouped together, because only one

mosquito oviposited at 25 days, and thus there was no way to generate a slope estimate.

Results

The winglengths of female Culex nigripalpus in this experiment ranged from 2.614mm to

3.227mm (Table 4-1). The range was less than that of a wild cohort captured in a lard can trap

(Chapter 2, Table 2-1), but the means were similar. The amount of hematin was also variable,









but again was less so than the mosquitoes captured in the study of wild Culex nigripalpus.

Fecundity ranged from 47-262 eggs with a mean of 171.

Each regression tested was significant at the a=0.05 level (Table 4-2). Simple linear

regressions of fecundity against hematin excreted, winglength and age were all significant, with

directions as predicted by theory. The slope of the simple linear regression of fecundity versus

age was -4.37, meaning that with every day of age added, mean fecundity would be expected to

decline by 4.37 eggs. This simple linear regression had the highest R2 Of any of the three

performed, yet even this was low compared to the multiple model.

In the multiple regression models, there was a significant interaction between age and

hematin after stepwise model reduction (Table 4-2). The standardized full model provides the

clearest picture of the relationships between the parameters, showing a negative correlation

between age and fecundity (Figure 4-1A), and positive correlations between fecundity and both

winglength and hematin (Figures 4.1B, C). On the other hand, the unstandardized model shows

a somewhat confusing result, namely that age has a positive slope, and is not a significant

predictor. This is largely due to the fact that models with higher-order terms (including

interactions) suffer from scales that have origins far from the centroid of the observed data

(Marquardt 1980). In the case of the unstandardized multiple regression model, each coefficient

is calculated holding every other factor at zero, which is not a value attained in this dataset. In

the case of the standardized model, each coefficient is calculated at the mean of each other

variable, providing a truer approximation of mean effect. For both qualitative and quantitative

interpretation, the standardized regression is preferable. The unstandardized form may be used

as a simple predictive equation.









Further analysis of the significant interaction by contrasting the slopes of the fecundity

versus hematin curves of the different age classes reveals a complex picture, where slopes for the

hematin fecundity regression were significantly different from zero at 5, 9, and 17 days, but not

at 13 days or the pooled values for 21-25 days (Table 4-3). Slope estimates, regardless of

significance, were positive until the group of mosquitoes aged 21-25 days (Figure 4-4).

Discussion

Age in insect vectors of disease is generally studied in order to appraise the potential of a

population to transmit disease. Because viral and other diseases require an extrinsic incubation

period in the vector in order for that vector to become capable of transmission, aging populations

are considered an important component in the disease transmission cycle. The effect of age on

fecundity of a vector population has been less often studied, but one may easily hypothesize

significant epidemiological consequences to alterations in reproductive rates of vectors. In

certain circumstances, large numbers of insect vectors of disease need to be present over a

transmission season in order for maj or outbreaks to occur (Mitchell et al. 1980, Day 2001).

To date there have been no published studies on the effect of age on fecundity of Culex

nigripalpus. The regression coefficient for the simple linear regression of age on fecundity in

this study is therefore the first documentation of an age effect on reproduction of this species.

The slope of the regression line (-4.374) compares well with a published study on a Vero Beach

strain of Culex quinquefasciatus which gave a slope of -4. 146 for the decrease in fecundity due

to age (Walter and Hacker 1974). Estimates for the rate of decrease in fecundity with age vary

by species and strain, and thus the value reported in this thesis is only representative of the

particular captive strain used. It may be that wild populations or different isolated captive strains

may respond to age in a different manner than reported here.









Culex nigripalpus is hypothesized to have substantial survival during drought conditions,

when oviposition sites are scarce (Day et al. 1990). It is also known that activity patterns of this

species are correlated with rainfall. If drought conditions prevail in the Hield, and oviposition and

bloodfeeding do not occur, the females that survive the drought will be older at reproduction than

if inclement conditions had not occurred. In this scenario, one could expect to see lower mean

clutch sizes and hence lower population growth in an aging population when compared to a

similar sized cohort of younger females.

This variability in reproductive output with age may be significant, but it can be argued

that the effect of mortality with age would far outstrip the reduction in fecundity. An estimate

for the daily rate of survival for this species was made with a recapture study of Fl adults

marked with radioactive phosphorus (Dow 1971). The estimate arrived at in this study was 81%,

meaning that the daily mortality rate is 19%. This estimate may be biased due to the fitness

consequences brought about by exposure to radiation, but it serves as a reminder that

survivorship of wild Culex can be quite low. Notwithstanding the sizable contribution of

mortality to overall Sitness parameters, the effect of age on fecundity would still be additive to

any population fitness reductions brought about by mortality.

Size of bloodmeal and size of adult female mosquitoes have often been reported as

significant predictors of fecundity. It has been reported that bloodmeal volume affects the first-

cycle fecundity of Culex nigripalpus if one measures this parameter by visual estimation (Edman

and Lynn 1975), but the methods used in that study did not allow calculation of a slope for that

relationship. This study first confirms, then extends the precision of this finding by providing a

means to quantify the amount of blood ingested and to analyze the relationship of this factor with

fecundity. Overall, the two simple regressions of winglength and hematin on fecundity were










significant, but when analyzed without considering the effect of age, they had relatively little

predictive power (Table 4-2).

The full model for prediction of fecundity from age, body size and bloodmeal size had

significantly greater predictive power than any of the simpler models, and fitting a model

including the significant interaction added approximately 4% to the predictive ability of the

equation (Data not shown). Before considering the interaction alone, it is important to first

discuss the relative merits of the two full models developed. In the unstandardized model, the

predictors are scaled in the familiar units of millimeters of winglength and micrograms of

hematin excreted. It would seem that using this model for analysis would be sufficient, and it is,

providing predicting fecundity from these factors is all that is required. If one wants to give the

relationship more consideration, such as what a typical change in either of these parameters does

to fecundity, the unstandardized model is no longer sufficient. In this dataset, there are no

deviations of a whole millimeter in winglength between any two female Culex nigripalpus. The

typical difference in winglength between any two randomly chosen females is only 0.13 mm

(Table 4-1). Using variables coded as Z-scores allows one to quickly determine what effect a

typical (1 standard deviation) change in a predictor will have on the response. Since all the

predictors are Z-scores, one can compare between typical variations and estimate the relative

importance of each predictor to changes in the response. Failure to standardize predictors also

has the unwelcome effect of obscuring significant effects of factors by making their partial

slopes seem small in comparison to others. This is only an effect of scaling, and disappears

when factors are properly standardized (Marquardt 1980). In addition, a regression using Z-

scores eliminates most of the "non-essential ill-conditioning" (multicollinearity) between

predictors that cause higher order models to behave erratically (Marquardt 1980).









An interaction between bloodmeal size and age has never been reported before in

mosquitoes, nor has such a possibility been considered. Therefore, the fact that a significant

interaction was found is both surprising and novel. However, the interaction was mainly due to

the last age group considered, and due to low survival, the power for estimating the effect was

low. Consideration of the interaction by grouping the mosquitoes into two arbitrary categories of

"young" and "old" demonstrates that in the older age groups, bloodmeal size has no significant

effect on fecundity, whereas in the younger age groups, bloodmeal size was a significant

predictor of fecundity. Possible mechanisms underlying this interaction are difficult to

hypothesize. Empirically, one can describe the interaction by saying that the reproductive

advantage conferred by a larger bloodmeal is only evident in younger mosquitoes. As

mosquitoes age, the advantages diminish until there is no significant relationship between size of

bloodmeal and fecundity. This type of interaction may only be evident in experimental settings

such as these, where mosquitoes are denied a bloodmeal until a given age, or this result may

generalize to aging mosquitoes regardless of their access to bloodmeals.

A possible physiological mechanism for this change in the relationship between bloodmeal

size and fecundity with age is an overall decline in the ability of an insect to synthesize protein as

it ages (Levenrook 1986). Protein enzymes are utilized in bloodmeal digestion (Clements 2000

c) as well as in nutrient transport to maturing oocytes (Clements 2000 a). If the efficiency of

these systems is substantially reduced, the overall advantage of larger bloodmeals may be

negated. A simple failure of the physiological systems underlying bloodmeal digestion and/or

nutrient transport would account for the observed interaction between bloodmeal size and age

predicting fecundity. It is interesting to note that in a study of Anopheles stephensi, Ferguson

and colleagues (2003) noted that the positive contribution of bloodmeal to fecundity disappeared









if the mosquito was infected with Pla~snodium chabaudi. Perhaps the state of senescence and the

state of Plasnzodium infection both affect digestion or reserve mobilization.

The novel finding of an interaction between bloodmeal size and age predicting fecundity

should be investigated further. Specifically, a study similar to this one, but utilizing a greater

number of age classes might better pinpoint the age at which increased bloodmeal size fails to

result in increased fecundity.

Another area that would be interesting to pursue would be to design an experiment where

multiple bloodmeals and ovipositions are provided and recorded. This would be a more complex

experiment, but it would allow for the quantifieation of the effect of age on total lifetime

fecundity in an experimental setting that is similar to a situation encountered by wild populations

of mosquitoes. It is unlikely that wild mosquitoes living to an age of 17 days or more would not

have had the opportunity to bloodfeed. Allowing for early and frequent bloodfeeding

opportunities in an experimental setting would help determine whether protein input from early

bloodmeals would better preserve digestive efficiency in later bloodmeals. It is at present

unclear whether protein resources acquired from bloodmeals can be used by a mosquito for its

own cellular maintenance and synthesis of enzymes. Mosquitoes can acquire extra-ovarian lipid

and carbohydrate from bloodmeals, but studies on utilization of bloodmeal protein suggest that

diversion of this resource to maternal reserves is minimal in Anophelines and non-existent in

Culicines (Clements 2000 b).

Conclusions

Good prediction of first-cycle fecundity of Culex nigripalpus may be achieved by using

bloodmeal size, body size and age as predictors, within the age range of 5 to 25 days of age. It is

likely that these factors are important in determining fecundity in wild populations of this species

as well. The contribution of bloodmeal size to fecundity declines with age of this mosquito.









This is a novel finding that bears closer scrutiny to determine if it holds across different age-

related nutritional regimes.













Table 4-1. Summary statistics for controlled and measured parameters used in regression
analyses.
Parameter
Stati stic Winglength Age Hematin Fecundity
Excreted
Minimum 2.61 5 1.12 47
Mean 2.86 11.8 11.63 171
Maximum 3.23 25 19.25 262
N 66 66 66 66
Standard 0.13 4.96 3.98 45
Deviation










Table 4-2. Summary of 5 regression calculations of various factors and combinations of factors
on fecundity.


*Effects not significant at the a=0.05 level discarded
t Indicates degrees of freedom for overall model F-test.


DFt T or F P R2


Model

Age


Parameter Coefficient


Std.
Error
12.72
0.9955
39.81


Intercept
Age
Model

Intercept
Winglength
Model

Intercept
Hematin
Model

Intercept
Winglength
Hematin
Age
Hematin*Age
Model

Intercept
Winglength
Hematin
Age
Hematin*Age
Model


222.48
-4.37


-96.03
93.27


17.4951
-4.3937
1,64 19.3

1,64 -0.8102
2.2546
5.083

1,64 7.5906
3.1364
9.837


<0.01
<0.01
<0.01

0.42
0.03
0.03

<0.01
<0.01
<0.01

0.2615
0.026
<0.01
0.48
0.01
<0.01


0.23


Winglength


118.5271
41.3709
43.71


0.07


Hem atin


122.86 16.1854
4.13 1.3176


0.13


Stepwise model,
unstandardized


-110.35
73.45
11.15
1.86
-0.58


171.43
9.63
16.97
-24.47
-11.54


97.3730
32.1597
2.9717
2.6099
0.2277
33.57

4.1363
4.2148
4.2325
4.2610
4.4951


-1.1333
2.2839
3.7529
0.7147
-2.5671


0.48


4,61


Stepwise model,
regressors
standardized


41.4451 <0.01 0.48


2.2839
4.0085
-5.7423
-2.5671


0.03
<0.01
<0.01
0.01
<0.01


4,61

















2so 'Fecundity=222.48 4.37 (age)

*
a,200



150 1 r
U)r e


S100*
C


L.



5 10 15 20 25

Age (Days)



Figure 4-1. Scatterplot with least squares regression line of the effect of age on fecundity of
Culex nigripalpus.





























60

























*



*
**
*
*
* -








Fecundity=-96.03 + 93.27(Winglength)


250




a, 200




150




S100

C



LL


2.5 2.6 2.7


2.8 2.9 3.0 3.1 3.2

Winglength (mm)


Figure 4-2. Scatterplot with least squares regression line of the effect of body size on fecundity
of Culex nigripalpus.


















Fecundity=122.86+ 4.13 (Hematin)*
250


*
20 *
*.
L e,
~3 1 ~e





LI. *
10





L.




3 8 13 18

Hematin (yg)



Figure 4-3. Scatterplot with least squares regression line of the effect of bloodmeal size on
fecundity of Cudex nigripalpus.

















250 ....... 9 days*
0 ---.--- 13 days
S--*-- 17 da s *m
20 -9- 21-25 days *



00 .*



oi 150
LI..:1---
O1
3 31
Heai odd(g

Fiue44 ctepo wt es qae ereso ie fteefeto loma ieo
feunit of Cre irplu.Nt hti h gs2-5dysoeetmta







Figue 4negSativealothough lath regres rgesion wasnot sinficn ath efeth a=.0 blodevel.s










Table 4-3. Summary of two ANCOVA analyses describing slope of the fecundity versus
standardized hematin curve in 5 age classes (slope at 25 days was not estimated due
to low sample size). Also shown are the slopes of the pooled "young" versus "old"
mosquitoes. Note that the slope in the "old" group is not significantly different from
zero.

Age (Days) Slope Estimate Std. Error n T P
5 6.002 2.6801 13 2.24 0.0291

9 8.425 1.951 17 4.32 <0.01

13 1.672 1.766 19 0.95 0.3479

17 5.259 2.156 12 2.44 0.0179

21-25 -7.958 4.373 5 -1.83 0.0732

Ages 5-13 4.42 1.41 49 3.13 <0.01

Ages 17-25 2.89 2.36 17 1.23 0.2250









CHAPTER 5
SENESCENCE AND FECUNDITY OF Culex quinquefasciatus

Introduction

Culex quinquefasciatus Say is a cosmopolitan mosquito species found in the tropics and

subtropics. It is a noted vector of diseases, including filariases (Ahid et al. 2000, Lima et al.

2003), protozoan parasites (van Riper III et al. 1986), and various arboviruses (Meyer et al. 1983,

Sardelis et al. 2001). Due to its status as a vector, much attention has been paid to the bionomics

of this species. Population density and dynamics of vectors such as this are of interest because

they are one element affecting the probability of disease transmission (Day 2005). Many factors

are known to affect the growth and maintenance of populations of Culex quinquefasciatus, at the

level of landscapes, microhabitats, and the physiology of individual mosquitoes. This study

focuses on some of the important predictors of reproductive output at the level of adult

physiology, as the ultimate determinant of population success is success of individuals (Briegel

2003).

Like all anautogenous mosquitoes, Culex quinquefasciatus depends on a bloodmeal for the

necessary nutrients (protein) to produce eggs. There have been many studies detailing the

relationships between bloodfeeding and reproductive output (Akoh et al. 1992, Hogg et al. 1996,

Roitberg and Gordon 2005). Studies on Culex quinquefasciatus have documented a positive

relationship between bloodmeal size and fecundity (Akoh et al. 1992, Lima et al. 2003). This

indicates that the size of the bloodmeal, in part, determines the number of eggs that can be laid.

Other factors are known to influence mosquito fecundity, including body size and general

reserves of the adult females (Briegel 2003). A portion of the protein reserves accumulated

during the larval period can be used for nourishment of eggs, and so better larval conditions

usually result in female mosquitoes capable of greater reproductive output. Body size is usually










positively correlated with larval nutrition, so it serves as an indicator of larval habitat quality,

and hence general reserves (Briegel 2003, Telang and Wells 2004). Many studies have shown a

positive relationship between female body size and egg production (Briegel 1990 a, 1990 b,

Akoh et al. 1992, Lyimo and Takken 1993, Armbruster and Hutchinson 2002, Lima et al. 2003).

Other members of this species complex, such as Culex pipiens molestus Forskal, a species

in the same complex of species (Vinogradova 2000), derive the entire protein input for egg

production from larval nutrition (are autogenous). It is not unreasonable to assume that a portion

of the protein required for egg production in the anautogneous members of this complex, such as

Culex quinquefasciatus is also derived from the larval stage.

A factor which is generally considered to lower reproductive output in many animal taxa is

age. In animals with high rates of daily mortality such as Culex mosquitoes (Dow 1971,

Elizondo-Quiroga et al. 2006), selection is likely to have favored high early-life reproduction, at

the expense of late-life fitness and reproductive capacity (Kirkwood and Rose 1991). This

follows from a theory known as the antagonistic pleiotropy hypothesis (Williams 1957). This

hypothesis states that mechanisms favoring early life fecundity may have deleterious effects in

later life. This is especially pronounced in animals with high per diem mortality such as

mosquitoes, since the numbers surviving to a late age are insignificant, as is the contribution of

these individuals to population growth. So far the physiological mechanisms underlying

fecundity reductions with age have not been determined.

A number of studies have examined declines in reproduction with age in mosquitoes, and

several of these (Walter and Hacker 1974, Gomez et al. 1977, Suleman and Reisen 1979, Akoh

et al. 1992) have detailed such declines in various strains of Culex quinquefasciatus. Most of the

studies just cited have not examined the role of other physiological parameters such as body size









and bloodmeal size, or have done so in a manner that does not include these other relevant

factors in a global model. Because of this, it is unclear what the relative contributions of each of

these factors to fecundity in an aging population truly are. The implication of this decline in

reproductive capacity with age for the population growth rate is unclear, but it may at times be

severe (Charlesworth 2000).

In the interest of developing predictive models for fecundity declines with age,

consideration of other factors such as bloodmeal size and body size may improve the robustness

and predictive qualities of the model. This study was designed to determine the contributions of

age, body size and bloodmeal size to the fecundity of Culex quinquefasciatus, considered

together in a multiple regression context.

Materials and Methods

Larval Rearing

Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing

approximately 700 ml of tapwater. Pans were set with 3 egg rafts each of colonized Culex

quinquefasciatus, from USDA ARS Gainesville FL, established 1995 (Allan et al. 2006). Food

was provided daily to each pan as 20 ml of slurry containing 20 mg/ml 1:1 Brewer' s yeast/liver

powder. This rearing regimen was chosen in an attempt to generate a range of sizes, while still

achieving relatively simultaneous emergence as adults.

Pupation

Pupae were placed in 500 ml cups in a large cage measuring 57 x 57 x 57cm and sugar was

provided to the emerging adults as 10% sucrose solution on cotton wicks. Adults were allowed

to emerge for 12 h following the emergence of the first female, whereupon the cups were

removed. The final density of mosquitoes in the cage was estimated to be about 700 females.










Temperature was 26.8 OC for the duration of the experiment. Relative humidity was 91.7%, and

the light cycle was 16:8 L:D.

Bloodfeeding

Beginning at 5 days post-eclosion, approximately 36 host-seeking females were removed

from the cage and bloodfed to repletion on a restrained chicken (University of Florida IUCUC #

D509). This was repeated at 4 day intervals until 6 groups were obtained. The age range thus

produced was 5-25 days post-eclosion.

Hematin Collection and Quantification

Bloodmeal size can be quantified by measuring the amount of hematin in the excreta.

Hematophagous arthropods void acid hematin as a byproduct of hemoglobin digestion. It has

been found that the quantity of hematin voided corresponds in a linear manner with the amount

of blood ingested (Briegel 1980, 2003). Determination of the amount of hematin in the excreta

thus provides an estimate of the relative amount of blood ingested. This method is easy to

implement on large numbers of mosquitoes and is minimally invasive.

Immediately following bloodfeeding, individual mosquitoes placed in separate 40 ml vials

(2.5 cm diameter, 9.5 cm deep) covered with screen. 10% sucrose was provided on small cotton

balls. Following a 4 day period for egg maturation, females were transferred to separate vials

(see below) for oviposition. The time required digestion and egg development for this species

has been found to be 2-3 days (Elizondo-Quiroga et al. 2006), and hence four days was found to

be a good balance between maximal survival in the vials and maximal ovarian development.

Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00 ml

1% LiCO3. The resulting solutions were decanted into spectrophotometric cuvettes and the

absorbance at 387 nm was read using a spectrophotometer. Absorbance readings were converted

to micrograms of hematin using a standard curve previously prepared (Appendix B).









Oviposition

After egg maturation, gravid females were transferred into a second set of 40ml vials

containing 4.0 ml of 10% (by volume) hay infusion in tap water for oviposition. The following

morning, egg rafts deposited were removed, placed on water under a microscope and

photographed at high magnification with a digital camera. Photographs of the egg rafts were

printed out using a standard laserj et printer and the number of eggs counted (method suggested

by A. Doumboya, personal communication, 2005).

Egg rafts were replaced in the vials and incubated at 27. 1C for 36 h to allow hatching.

Hatched larvae were filtered onto white filter paper and counted. Percentage hatch was

calculated as the number of larvae hatched divided by the number of eggs multiplied by 100. It

is unknown whether the handling of the egg rafts affected percentage hatch. Delayed hatching is

not known from subgenus Culex mosquitoes, but has been observed in the subgenus

M~elan2oconion (Hair 1968).

Winglengths

Following oviposition, females were removed, killed, and their wings excised and mounted

on slides for measurement. Abdomens were dissected to count retained eggs. The wings on the

microscope slides were photographed with a standard size reference (a length of steel measured

to the thousandth of a millimeter with a caliper). The photographs were opened in SigmaScan

Pro 5 (Systat Software, Inc., Point Richmond, CA), calibrated for size, and measured from the

alular notch to the distal end of R2, excluding fringe hairs (Packer and Corbet 1989). It was

decided to use the distal end of R2 aS a measurement point as it is an unambiguous standard

landmark. Other studies have used the distal end of the wing, or some kind of other subj ective

measure of the maximal distance (Packer and Corbet 1989, Lima et al. 2003). While others have

suggested transforming the winglengths thus obtained by cubing the linear measure (Briegel









1990 a), the recommendations of Siegel were followed here (Siegel et al. 1992), thus

winglengths were not transformed.

Analyses

Regression analyses were conducted using S-Plus@ 7.0 for Windows@ (Insightful 2005).

Analysis of Covariance was conducted using SAS 9.0 (SAS Institute). Fecundity was scored as

the number of eggs in the egg raft plus the number of retained eggs in the killed females,

provided these numbered fewer than 50. If females retained more than 50 eggs, they were

discarded from the analysis, as they became difficult to count in high numbers. Only two

females in this study retained more than 50 eggs.

Simple linear regressions of each predictor on fecundity were performed to compare with

published results of these predictors on fecundity for Culex quinquefasciatus females.

A multiple linear regression was performed on untransformed variables regressing

fecundity on body size, bloodmeal size and age. Non-significant interactions and main effects

were discarded in a stepwise fashion. Normality was checked with a Kolmogorov-Smirnov Test

of Composite Normality, and homoscedasticity with a plot of residuals versus fitted values.

A multiple regression of fecundity on the standardized values (Z-scores) of the predictors

was also performed. Z-scores were calculated by subtracting the mean value of a regressor from

each observation, then dividing this by the standard deviation of the regressor. This produces

mean values of zero and standard deviations of one. When used in multiple regression analysis,

it allows a more standard interpretation of slope values, i.e. It apportions mean changes in

response due to predictor variations of one standard deviation (Marquardt 1980). Doing so

allows one to order the predictors in terms of influence on a common scale.










Significant interactions with age were explored with an ANCOVA testing for differences

in slope of hematin or winglength on fecundity associated with the different age groups.

A regression of arcsine square root transformed percentage hatch on age was performed,

but every transformation of the data failed to produce normal residuals. Further analysis of this

relationship was conducted with a Kruskal-Wallis test (Milton 1992). A regression of raw

percentage hatch against age was performed to get a rough estimate of the effect of age on %age

hatch.

Results

Winglength of female Culex quinquefasciatus ranged from 2.77 to 3.39 mm (Table 5-1).

Amounts of hematin excreted and fecundity also varied. (Table 5-1).

Significant linear regressions were found for all three factors, as well as the multiple

regressions (Table 5-2, Figures 5-1, 5-2, 5-3). The fit of the full multiple model was

considerably better than for any of the one-factor linear regressions. All individual parameters

influenced fecundity with directions predicted by previous work, except for the unstandardized

multiple model (Table 5-2). The standardized multiple model indicates the expected direction of

influence of these parameters on fecundity (hematin positive, winglength positive, age negative).

In the formation of the multiple models, it was discovered that there was a significant

interaction between age and body size predicting fecundity. Analysis of this interaction by

ANCOVA (slopes for winglength at varying levels of age) showed that in the younger age

classes, slopes were positive and significant, (Table 5-3), but after 17 days of age, slopes were

not significantly different from zero. For illustration, significant slopes were grouped, and slopes

not different from zero were grouped (Fig. 5-4).









The Kruskall-Wallis test showed significant differences in %age hatch between the six age

groups (Chi-square = 20.7657, df = 5, p-value = 0.0009), but an examination of the scatterplot of

percentage hatch versus age failed to show a reliable trend with age (Figure 5-5).

Discussion

As expected, both bloodmeal size and body size had a positive influence on fecundity. The

degree of dependence on these two factors considered alone is relatively low, but taken together

with age, a very precise model explaining most of the variance in fecundity was achieved. Age

was the factor that explained the greatest variance in fecundity, either as a single factor in the

simple linear regressions, or as a factor in the standardized multiple regression model.

The decline in fecundity with age had the greatest influence on fecundity in the

standardized multiple model, indicating that a standard deviation in age has a greater effect than

a standard deviation in bloodmeal size or body size. That there is a noticeable decline in

fecundity with age over a span of 25 days post-eclosion accords with one of the expectations of

the antagonistic pleiotropy hypothesis of aging, namely that age-associated fitness effects are

expected to be large in taxa experiencing high per-diem mortality (Williams 1957).

Consideration of the two full models illustrates the potentially confusing nature of

unstandardized factors used as predictors in multiple regression models. The unstandardized

form seems to qualitatively differ with respect to the expected sign (+/-) for age. The

explanation for this is that factors were measured on different scales (Marquardt 1980). Each

regression would perform identically when used as a predictive equation, but more intensive

analysis is better accomplished by examining the model with standardized regressors.

In the standardized model, each predictor accords with the expected direction of influence

on fecundity: i.e. Negative for age, positive for bloodmeal size and body size. Since the

predictors are considered at equivalent intervals, one can judge the relative contributions of each









to the overall model. On the whole, age had the strongest influence on fecundity, followed by

bloodmeal size and body size.

The estimate for the decline in fecundity associated with age in the unstandardized simple

linear regression of age on fecundity is similar to that reported for a wild Vero Beach strain of

this species (Walter and Hacker 1974), but greater than those reported for Asian populations of

this species (Walter and Hacker 1974, Suleman and Reisen 1979). The influence of age on

fecundity may also be modified by temperature and humidity, so direct comparisons between

these studies are difficult to make, as these other studies held the animals under different

conditions..

It should be noted that the effect of age-related declines in fecundity are only one measure

of the impact of age on population growth in an age-structured population. Mosquitoes have

high rates of daily mortality (Dow 1971, Reisen et al. 1991), and this is likely to account for a

greater decline in fitness at the population level than changes in fecundity (Dye 1984).

Nonetheless, the effect of these age-related fecundity declines is substantial and should be

incorporated into the construction of models of age-structured population growth wherever

possible (Gotelli 1998).

The interaction between adult body size and age accounted for a portion of the variance

and increased the R2 Of the model moderately. It also led to a more careful consideration of the

effect of body size across the different age groups. The effect of body size on fecundity was not

constant: in the younger groups, it was significant and positive whereas in the older groups it had

no significant influence. A possible explanation would be that if mobilization of general protein

reserves is interrupted as a consequence of age, or if most of these reserves have disappeared,

then the only significant factor predicting fecundity would be bloodmeal size.










Briegel demonstrated with Aedes aegypti the somewhat surprising fact that larger females

invested progressively less lipid into their yolk mass with increasing gonotrophic age, despite

having access to sucrose from which to synthesize lipids. This was not seen in smaller females

(Briegel et al. 2002). In fact, in both groups, total lipid increased with age. Associated with this

finding was an observation that the slope of the fecundity versus age curve was steeper for larger

females than it was for smaller females, indicating that proportionate to their body size, larger

females tended to produce less eggs in later ovipositions than in earlier ones. If such a dynamic

also exists in Culex quinquefasciatus, then this could show up as an interaction between age and

body size predicting fecundity.

Another possible explanation for this might be a decline in protein reserves brought about

by the sugar-only diet of the females prior to bloodfeeding (Lang et al. 1965). If protein reserves

are totally exhausted, then they cannot positively influence fecundity. If we also postulate that

there is no or low correlation between body size and bloodmeal size, then any positive effect of

body size on fecundity would disappear when general reserves are exhausted. In this dataset,

only 12.9% of the variance in bloodmeal size can be apportioned to differences in body size

(data not shown).

The relationship between general protein reserves and fecundity is best seen with

autogenous mosquitoes, as protein for the first egg batch is derived entirely from general reserves

(Telang and Wells 2004). Because in this study the correlation between body size and

bloodmeal size is low (larger females did not take significantly larger bloodmeals), there is no

discernable advantage to being larger.

This is the first time that an interaction between adult mosquito body size and age

predicting fecundity has been reported, and it should stimulate further research into the









possibility of interactions with age in predictive models of mosquito fecundity. Particular

consideration should be given to generating a wide variety of sizes of mosquito, as well as

perhaps greater diversity of age groups. This may give a better idea of the timing of any

interaction effects. Other experiments should also be performed testing whether this type of

interaction between body size and age is evident in females given multiple bloodmeals.










Table 5-1. Summary statistics for factors and responses used in regression models.
Stati stic Hematin Winglength Age Fecundity (# of Retained percentage
(Clg) (mm) eggs in raft) eggs hatch
Min 7.47 2.77 5 43 0 1.15
Mean 16.87 3.06 12.14 151 1.27 81.47
Max 29.39 3.39 25 231 35 100
N 130 130 130 130 130 130
Std. 4.04 0.11 6.98 44.17 4.58 18.58
Deviation










Table 5-2. Summaries of five separate regression equations predicting fecundity of Culex
quinquefasciatus fecundity from individual factors and combinations of factors.


DFt T or F P R2


Regressors


Parameter

Intercept
Age
Model

Intercept
Hem atin
Model

Intercept
Winglength
Model


Coefficient Std.
Error
222.61 6.5136
-5.75 0.4901
30.24

72.85 14.79
4.74 0.85


Age


34.18
-11.72
1,128 137.40

4.93
5.56
1,128 30.95


<0.01
<0.01
<0.01

<0.01
<0.01
<0.01

0.23
<0.01
<0.01

<0.01
<0.01
<0.01
<0.05
<0.01
<0.01

<0.01
<0.01
<0.01
<0.01
<0.01
<0.01


0.52




0.19


Hem atin


Winglength


-123.88
90.31


-409.13
3.60
187.03
24.20
-9.81


153.50
14.54
7.56


102.71
33.49


137.48
0.54
45.59
10.79
3.53


-1.21
2.70
7.27


0.05


1,128


Full model,
unstandardized


Intercept
Hem atin
Winglength
Age
Winglength*Age
Model

Intercept
Hem atin
Winglength
Age
Winglength*Age
Model


-2.98
6.71
4.10
2.24
-2.78
4,125 85.04

76.13
6.71
3.47
-15.64
-2.78
4.125 85.04


0.73


Full model,
standardized


2.02
2.17
2.18


0.73


-31.90 2.04
-5.94 2.14


*The predictors and interactions that were not significant at the a=0.05 level were removed as
part of a stepwise model reduction.
? Indicates degrees of freedom for overall model F-test.



















~3200



150




w100




a, so
u..


* *


Fecundity= 222.61 5.74(Age)


10 15 2

Age (Days post-eclosion)


Figure 5-1. Scatterplot of fecundity versus unstandardized age at bloodfeeding for Culex
quinquefasciatus.






















e
*


TD 200





150





a100




O 50
L.


*


*



*

*


*C


** *


**


Fecundity=72.85 + 4.74 (Hematin)


15 20

Hematin Excreted ( pg)


Figure 5-2. Scatterplot of fecundity versus unstandardized bloodmeal size for Cudex
quinquefasciatus.


















e *

*
99





150



*


o *~ (


u. *


Fecu ndity=-1 23. 88 + 90.31 (Wi ng length)



2.7 2.9 3.1 3.3

Winglength (mm)

Figure 5-3. Scatterplot of fecundity versus unstandardized body size for Culex quinquefasciatus.



































80









































O
2.7 2.9 3.1 3.3

Winglength (mm)
Figure 5-4. Scatterplot of unstandardized winglength versus fecundity showing the response
between ages 5- 13 days (solid line, circles) and 17-25 days (broken line, triangles).


Table 5-3. ANCOVA analysis describing slope of the fecundity versus winglength relationship
for 6 Culex quinquefasciatus age classes. Also shown are the slopes of the pooled

"young" versus "old" mosquitoes. Note that in both cases, after 13 days, the slopes
are not significantly different from zero.

Age (Days) Parameter Estimate Std. Error n T P
5 Winglength 151.043 41.488 27 3.64 <0.01


9 Winglength 115.007 45.116 32 2.55 0.0121


13 Winglength 175.785 36.796 32 4.78 <0.01


17 Winglength 10.516 48.545 24 0.22 0.8289


21 Winglength -109.853 103.163 11 -1.06 0.2891


25 Winglength 54.767 129.188 4 0.42 0.6724


Ages 5-13 Winglength 131.773 27.743 91 4.75 <0.01


Ages 17-25 Winglength -6.317 49.090 39 -0.13 0.8978


250




5 200




S150




100




0 50
L.







~
+

~
V
R
+

-,,

~~a, Y
a a
4

a~ 1l
1---11-I I '1111- 1 I I.

a














100- *

80 I
I I *
I

r* *


2 0-








5 10 15 20 25
Age (Days)

Figure 5-5. Scatterplot of Percentage hatch versus age for Culex quinquefasciatus.









CHAPTER 6
DISCUSSION

Fecundity is an important parameter in the overall fitness of female organisms, as

well as males, albeit indirectly. At the population level, the mean number of offspring

produced by females has an effect on the growth characteristics of the population.

Anautogenous mosquitoes, like many insects, are capable of producing a great

number of eggs in their lifetime. This reproductive output is predicated upon several

factors, such as larval nutrition (Akoh et al. 1992), adult nutrient acquisition

(bloodfeeding) (Akoh et al. 1992, Briegel 2003), and survival. Another factor which is

known to affect reproductive rate in many organisms is age (Akoh et al. 1992, Mahmood

et al. 2004). The gradual breakdown in physiological competence, and associated

decreases in performance as an organism ages is referred to as senescence.

The studies described in this work presented here examined the role of bloodmeal

size, body size and age in determining the fecundity of Culex nigripalpus and Culex

quinquefasciatus. Body size served as the proxy for larval nutrient acquisition, since the

two factors have been found to be correlated (Akoh et al. 1992, Koella and Offenberg

1999).

Larval Nutrition

The effects of mosquito larval nutrition on reproduction have been well studied in

autogenous species. In autogenous species, the entire protein input into the yolk of the

eggs is of larval origin (Clements 2000 a) and no bloodfeeding is required, at least for the

first gonotrophic cycle. Lipid and carbohydrate may be larvally-derived, or may be

acquired from sugar feeding.









Anautogenous species (species requiring blood for vitellogenesis) likewise realize

benefits from the larval stage, both from the general reserves acquired (Briegel 2003), and

also because greater larval nutrition correlates with greater adult body size, which in turn

is correlated with a larger number of ovarioles in the ovaries (Clements 2000 c). Larger

mosquitoes are also able to take larger bloodmeals (though the strength of this

relationship varies). This means that larger mosquitoes are capable of greater protein

acquisition. The effects of larval nutrition are seen as a simple correlation between body

size and fecundity in the species in question (Briegel 1990 a, Akoh et al. 1992, Lyimo

and Takken 1993, Armbruster and Hutchinson 2002, Lima et al. 2003).

In the series of experiments reported here, there was consistently a significant

effect of body size on fecundity, indicating that larval nutrient reserves and adult body

size are important predictors of fecundity in Culex nigripalpus and Culex

quinquefasciatus. This result agrees with results obtained for many mosquito species

(Nasci 1986, Briegel 1990 b, Akoh et al. 1992, Lyimo and Koella 1992, Lyimo and

Takken 1993, Xue et al. 1995, Armbruster and Hutchinson 2002, Lima et al. 2003).

Bloodmeal Size

The size and quality of the bloodmeal is another important determinant of fecundity

in anautogenous mosquitoes. The bloodmeal provides the only protein input for egg

production aside from protein acquired in the larval stage. Blood also furnishes

substantial amounts of lipid and carbohydrate, which are used in vitellogenesis.

In most studies to date, the size of the bloodmeal has been shown to be positively

correlated with the number of eggs produced (Cochrane 1972, Edman and Lynn 1975,

Clements 2000 a, Briegel 2003). This has been shown to be the case in both Culex

nigripalpus (Edman and Lynn 1975), and Culex quinquefasciatus (Akoh et al. 1992).









The results of the experiments reported herein confirm these results. They also

provide a precise quantification of the number of eggs produced for a given amount of

voided hematin. The strength of these relationships varied by species, and also with age.

This provides a baseline estimate of the strength of the dependence of fecundity on

bloodmeal size in these two species of mosquitoes, as well as how bloodmeal size

combines with other factors to influence fecundity. This knowledge can be used to better

understand the role of life history events (such as bloodfeeding and larval feeding)

shaping the habits of these species.

Age

The process by which the health and vitality of an organism diminish with age is

known as senescence, and it is found to varying degrees in many taxa. What is less well

understood are the processes which contribute to this decline in faculties as an organism

ages.

There are some evolutionary theories regarding the origin and maintenance of

senescence in natural populations (Gavrilov and Gavrilova, 2002). The major predictions

of these theories are that short lived organisms are expected to experience the negative

physiological consequences of aging more severely and at an earlier age than longer lived

organisms, due to the weakness of selection against late-life fitness declines.

The phenomenon of senescence has implications not only for the aging individual,

but also collectively for the population as a whole, when cohorts of organisms become

old-biased. Senescence is a plastic trait under natural selection, and can be influenced by

artificial selection in relatively few generations (Stearns et al. 2000 a). This helps to

confirm an evolutionary explanation for senescence.










Mosquitoes are known to have high extrinsic mortality (Dow 1971, Reisen et al.

1991, Elizondo-Quiroga et al. 2006), and as such have likely been selected for high early-

life fecundity (an extrapolation from the theories mentioned above). This means that

reproductive capacity should reach its maximum shortly after the animal reaches sexual

maturity (4 days post-eclosion for many mosquito species) and decline thereafter. This

has been demonstrated in several mosquito species (Jalil 1974, Walter and Hacker 1974,

Gomez et al. 1977, Suleman and Reisen 1979, Akoh et al. 1992, Mahmood et al. 2004),

but the rate at which fecundity declines can be specific to the species or the geographic

strain of the species in question (Walter and Hacker 1974).

In disease vectors such as mosquitoes, the decline in fecundity with age may have

epidemiologic significance if it serves to decrease the abundance of vectors available for

enzootic or epidemic disease transmission over the course of a transmission season.

Arbovirus transmission is in part dependent, in part, on vector abundance, and if

abundance is not sufficient, transmission is unlikely (Mitchell et al. 1980, 1983). It

should be noted, however, that small populations of aged mosquitoes can sometimes

sustain epidemic transmission (Day 2001).

The experiments reported herein provide estimates as to the effect of age on the

fecundity of Culex nigripalpus and Culex quinquefasciatus from Florida. The

demonstration of an age-related fecundity decline in Culex nigripalpus is the first such

reported for this species.

The aging experiments in this work were conducted in a manner that did not allow

repeated bloodfeeding, rather the bloodmeals given represented the first that any female

had in her lifetime. It is not clear from these results whether the effects of senescence










might have been ameliorated had the females had access to blood on a regular basis.

Nonetheless, the slopes of the age related fecundity declines are similar to those of other

Culex species studied (Walter and Hacker 1974, Mahmood et al. 2004).

Interactions

An interesting and novel result to emerge from this work is the demonstration of

interactions between bloodmeal size (Chapter 4), and body size (Chapter 5) with age.

This is the first time such a phenomenon has been investigated, and bears closer

consideration.

In the case of the interaction between bloodmeal size and age predicting fecundity

in Culex nigripalpus, the interaction was in large part due to the response of fecundity to

bloodmeal size in the last two age groups (20 and 25 days old). The sample size for

estimating this effect was low, but the slope estimate was large and negative. A possible

explanation for the result is that digestive or vitellogenic processes are impaired in older

females.

In the case of the interaction between body size and age in Culex quinquefasciatus,

the result was more robust due to a larger sample size. A simple physiological

explanation suggests itself: that general reserves of protein are depleted as the mosquito

ages and thus at older ages, increased body size fails to result in increased fecundity.

Another explanation might be offered by the finding of Briegel, working with

Aedes aegypti: that larger females invest progressively less lipid into cogenesis as they

age, while smaller females do not (Briegel et al. 2002). Differentiating between these two

hypotheses could only be accomplished by careful biochemical analyses of the

mosquitoes as they age.









It is unknown if this interaction effect would disappear if repeated bloodmeals had

been made available. If bloodfeeding had been allowed at regular intervals, perhaps

metabolically-depleted general protein reserves would be replenished, although in

general, bloodmeal protein is generally only used for egg production (Briegel 2003).

Future Research

The results of these experiments suggest some new avenues and techniques for

research into factors governing fecundity in mosquitoes.

The interaction terms discovered in Chapters 4 and 5 suggest that there may be

some important physiological changes in blood digestion or levels of general reserves that

occur in Culex mosquitoes with age. With reference to the interaction between body size

and age predicting fecundity in Culex quinquefasciatus, the role of general protein in egg

production should be examined. It is known that the entire protein input necessary for

first-cycle egg production in autogenous mosquitoes, including Culex pipiens molestus

Forskal, are derived from larval nutrition (Clements 2000 c). It is not therefore

unreasonable to suspect that general protein also contributes to reproduction in

anautogenous species such as Culex quinquefasciatus. If so, the depletion of such protein

reserves in aging adult mosquitoes could be examined. Different taxa and populations

could then be compared in order to determine the relative contributions of general and

bloodmeal-derived protein inputs used for reproduction. This would be of interest

because it would allow a partitioning of the effects of larval and adult nutrition on

lifetime fitness.

With reference to the interaction between bloodmeal size and age predicting

fecundity detected in Culex nigripalpus (Chapter 4), the functioning of digestive and

oogenic processes should be examined in aging mosquitoes. If these become










compromised in some way, then the interaction could be explained, and a mechanism of

senescence in this species could be elucidated.

Aging and senescence are not biologically Eixed traits, but rather have been shown

to be modified by selection (Stearns et al. 2000 b). It is entirely reasonable to suspect

that different species and different populations within a species have been selected for

different lifespans. This is perhaps one reason why the decline in fecundity with age has

been shown to differ markedly between populations of mosquitoes (Walter and Hacker

1974). For this reason, life history traits such as senescence need to be examined with

reference to a particular population. For example, if one were to incorporate an age-

related fecundity reduction into an age-structured population growth model, one would

need to be careful to have parameter estimates derived from the population under

consideration. If not, then the model and the Hield reality may differ greatly.

One methodological consideration that these results suggest is that multiple factors

should be included together in models predicting fecundity as well as experiments to

determine factors affecting fecundity. This allows for greater accuracy in prediction, and

also allows some comparison of the relative importance of the factors under

consideration. It should also be obvious that the standardization of parameters improves

the interpretability of regression models, especially when higher order terms or

interactions are present. The dangers of working with unstandardized predictors in

regression models have been stated before (Marquardt 1980), namely that significant

effects could be unjustifiably discarded.

From a theoretical standpoint, it is obvious that possible interactions should be

sought out in multiple models, and if present analyzed further. An interaction indicates









that the response to one factor changes at different levels of the interacting factor. This

may be seen as a complication, but in reality, inclusion of significant interaction terms

can often better describe the relationship between multiple predictors and a response.

The results of these experiments shed some new light onto the interplay of factors

affecting fecundity in two Culex species. Continued investigation into the factors

governing reproductive success of mosquitoes will provide interesting insights into the

natural history of these fascinating and important insects.










APPENDIX A
WINTG MEASUREMENT VALIDATION

Objective

To test and validate a method using photographs and image analysis software for precisely

measuring mosquito winglength.

Methods

Ten pieces of aluminum from a soft drink can were cut to random sizes and measured to

the nearest thousandth of a millimeter with a caliper. The pieces were numbered and

measurements were recorded. The metal pieces were then photographed adj acent to a 6.758mm

reference. The photographs were imported into SigmaScan (Systat Software Inc. Richmond CA)

and calibrated using the SigmaScan 2-point recalibration function. The metal pieces were then

measured using the SigmaScan default 2-point measurement system in the program and the

measurements for each were stored.

A linear regression predicting callipered measurement from the SigmaScan measurement

was performed using S-Plus (Insightful Corp. Seattle WA).

Results

The linear regression was highly significant (Fig. A-1) and had an intercept near zero and a

slope that approached one (Table A-1). The intercept was not significant, and thus cannot be

distinguished from zero.

Conclusion

This method of measurement is accurate, precise and easy to use. It can easily be used to

measure between two distinct landmarks on a mosquito wing mounted on a microscope slide

(Fig. 2-1).



















S12-





-8-








0



0 2 4 6 8 10 12 14
SigmaScan M measure me nt (mm)

Figure A-1. Regression plot showing relationship between caliper measurement and SigmaScan
measurement.










Table A-1. Results of linear regression analysis of callipered length against length measured in
SigmaScan.
Parameter Estimate Standard Error P
Intercept 0.0542 0.0693 0.4573
Slope 1.0021 0.0101 <0.0001









APPENDIX B
HEMATINT STANDARD CURVE

Introduction

Hematophagous insects void several products in their feces following a blood meal. These

products include uric acid derived from tubal fluids (Briegel 1980, Clements 2000 d), as well as

products of blood digestion following a bloodmeal (Clements 2000 d). One of these products is

hematin, a degraded digestion product of hemoglobin. It was suggested by Briegel (1980) that

quantifying the hematin content of excrement might allow one to estimate the relative size of a

mosquito bloodmeal. The method of quantification involves reading the absorbance of the

sample at a wavelength of 387 nm, the absorption peak for hematin. Figure B-1 illustrates a

sample spectral scan of Culex quinquefasciatus excrement showing the absorption peak

corresponding to the maximum for hematin.

Many authors have used this approach successfully to quantify bloodmeals their

experimental animals consume (Briegel 1980, Mitchell and Briegel 1989, Briegel 1990 a, Hogg

et al. 1996, Ferguson et al. 2003). The chief advantage of this method of bloodmeal

quantification is that it does not involve intrusive and laborious methods of anesthetic use on

experimental animals and weighing of animals before and after bloodfeeding.



Objective

To develop a regression model that accurately calculates the amount of hematin in

mosquito excrement, as a means of quantifying the relative size of the bloodmeal.

Methods

Three stock solutions of 1.00 mg/ml porcine hematin (MP Biomedicals, Irvine CA) in 1%

lithium carbonate were prepared using an analytical balance. From these stock solutions, 14










dilutions were prepared, from 0-26Clg/ml at intervals of 2.00 Clg/ml. The absorbance of these

solutions at 387nm (the peak absorbance for hematin) was measured and the data recorded.

Absorbance at 387nm was regressed on hematin concentration using a linear regression in

S-Pluse 7.0 for Windowse (Insightful Corp.). The slope estimate for this regression was used in

all calculations of hematin concentration used in this Thesis.

Results

A regression equation was constructed to predict hematin mass from absorbance.

Conclusion

This technique allows one to quantify the hematin present in a sample of excreta, thereby

providing an indirect estimate of mosquito bloodmeal sizes.























10 0 0 0 0 0
Waeegh(m
FiueB1 Eapeo seta ca fClxqunufsiau xrmnt oepa
corsodn omxmlasrac o eai t37m






































h 387



Figure B-2. Standard curve used for calculating amount of hematin in mosquito excreta. R2 foT
this regression was 0.9969, F=12830 on 1 and 40 degrees of freedom. N=42


Table B-1. Results of the regression analysis for absorbance versus concentration. The slope
estimate was used to calculate all hematin quantities in this work.
Parameter Estimate Standard Error P
Intercept 0.0556 0.1346 0.6817
Slope 14.366 0.1268 <0.01


25

-
S20

-
ct15


o 10




"I
0









APPENDIX C
MOSQUITO WINGLENGTH VS. WEIGHT REGRESSIONS

Introduction

Several studies have show that a repeatable, meaningful measure of overall size of a

mosquito is the length of the wing (Briegel 1990 a, Nasci 1990, Siegel et al. 1992, Hogg et al.

1996, Koella and Offenberg 1999, Armbruster and Hutchinson 2002, Lima et al. 2003). This is

usually accomplished by measuring the wing from the alular notch to the wingtip, excluding

fringe hairs.

Objective

To determine the relationships between general weights of Culex quinquefasciatus and Cx.

nigripalpus females and their winglengths. The purpose of this was to find a proxy measurement

for female size for use in regression models predicting the fecundity of these species.

Materials and Methods

Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing

approximately 700 ml of tap water. Pans were set with 3 egg rafts each of colonized Culex

quinquefasciatus or Culex nigripalpus. Food was provided daily to each pan as 20 ml of slurry

containing 20 mg/ml 1:1 Brewer's yeast/liver powder.

Forty-two female pupae of Culex quinquefasciatus and 43 female pupae of Culex

nigripalpus were removed individually, and dabbed dry with a paper towel. Pupae were weighed

on a Cahn Millibalance Model 7500 to the nearest 100th of a milligram. Following weighing,

pupae were placed in individual 40 mm vials with 4 ml tapwater and allowed to emerge.

Following emergence, the mosquitoes were killed and their wings removed and adhered to a

microscope slide with clear double-sided tape. A coverslip was placed over the wings to prevent

damage. Wings were photographed next to standard size reference and the distance between the









alular notch and the end of wing vein R2 (excluding fringe setae) was measured using SigmaScan

(Appendix 1).

Data were entered into S-Plus and linear regressions performed predicting mass from

winglength.

Significant linear relationships were found between pupal mass and winglength in both

species (Table B-1, Figs. B-1, B-2). Slopes and intercepts for both regressions were quite similar,

with that for Culex quinquefasciatus being slightly higher.

Table C-1. Results of two regression analyses predicting female mass from winglength. All
regressions were significant at the a=0.05 level.
Species Parameter Estimate Standard df T P R2
Error
C'x. Intercept -4.17 1.77 1,40 -2.36 0.02 0.34
quinquefasciatus Winglength 2.53 .56 4.54 <0.01
Cx. nigripalpus Intercept -4.12 .98 1,41 -4.21 <0.01 0.57
Winglength 2.33 .32 7.37 <0.01

Conclusion

Winglength is an adequate predictor of pupal mass in Culex quinquefasciatus and Culex

nigripalpus.



















4.5





E 4.0 *




3.5 *





3.0 *


2.9 3.0 3.1 3.2 3.3

Winglength (mm)

Figure C-1. Linear regression of winglength on wet mass of pupae of female Culex
quinquefasciatus n=42































100




Full Text

PAGE 1

1 SENESCENCE AND OTHER FACTORS AFFE CT FECUNDITY IN TWO SPECIES OF Culex MOSQUITOES (DIPTERA: CULICIDAE) By SEAN MICHAEL MCCANN 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 2006

PAGE 2

2 Copyright 2006 by Sean Michael McCann

PAGE 3

3 To my parents.

PAGE 4

4 ACKNOWLEDGMENTS I gratefully acknowledge the assistance of my supervisory committee, the laboratory staff, and various faculty around FMEL for their as sistance with time, material or advice.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............11 CHAPTER 1 INTRODUCTION................................................................................................................... ...13 Mosquitoes..................................................................................................................... .........13 Taxonomy....................................................................................................................... .13 Distribution and Notes on Natural History......................................................................13 Vector Associations.........................................................................................................14 Age............................................................................................................................ ......15 Bloodmeal Size................................................................................................................. ......17 Bloodmeal Source............................................................................................................... ....17 Body Size and Teneral Reserves............................................................................................18 Multiple Factors............................................................................................................... .......19 Summary........................................................................................................................ .........19 2 EFFECTS OF BLOODMEAL SIZE AND BODY SIZE ON FECUNDITY OF WILD Culex nigripalpus ....................................................................................................................23 Introduction................................................................................................................... ..........23 Materials and Methods.......................................................................................................... .24 Trapping....................................................................................................................... ...24 Hematin Collection and Analysis....................................................................................24 Oviposition.................................................................................................................... ..25 Winglengths.................................................................................................................... .25 Analyses....................................................................................................................... ...26 Results........................................................................................................................ .............27 Path Analysis.................................................................................................................. .27 Discussion..................................................................................................................... ..........28 3 INFLUENCE OF BLOODMEAL SIZE AND BODY SIZE ON THE FECUNDITY OF CAPTIVE Culex quinquefasciatus .........................................................................................35 Introduction................................................................................................................... ..........35 Materials and Methods.......................................................................................................... .36 Larval Rearing.................................................................................................................36 Pupation and Bloodfeeding.............................................................................................36

PAGE 6

6 Hematin Collection and Quantification...........................................................................37 Oviposition.................................................................................................................... ..37 Winglengths.................................................................................................................... .38 Analyses....................................................................................................................... ...38 Results........................................................................................................................ .............39 Discussion..................................................................................................................... ..........40 4 Culex nigripalpus AGING AND FECUNDITY.........................................................................45 Introduction................................................................................................................... ..........45 Materials and Methods.......................................................................................................... .46 Larval Rearing..........................................................................................................46 Pupation....................................................................................................................46 Bloodfeeding............................................................................................................47 Termination of Study...............................................................................................47 Hematin Collection and Quantification...........................................................................47 Oviposition.................................................................................................................... ..48 Winglengths.................................................................................................................... .49 Analyses...................................................................................................................49 Results........................................................................................................................ .............50 Discussion..................................................................................................................... ..........52 Conclusions.................................................................................................................... .........56 5 SENESCENCE AND FECUNDITY OF Culex quinquefasciatus .............................................65 Introduction................................................................................................................... ..........65 Materials and Methods.......................................................................................................... .67 Larval Rearing.................................................................................................................67 Pupation....................................................................................................................... ....67 Bloodfeeding...................................................................................................................68 Hematin Collection and Quantification...........................................................................68 Oviposition.................................................................................................................... ..69 Winglengths.................................................................................................................... .69 Analyses....................................................................................................................... ...70 Results........................................................................................................................ .............71 Discussion..................................................................................................................... ..........72 6 DISCUSSION..................................................................................................................... ........83 Larval Nutrition............................................................................................................... .......83 Bloodmeal Size................................................................................................................. ......84 Age............................................................................................................................ ..............85 Interactions................................................................................................................... ..........87 Future Research................................................................................................................ ......88

PAGE 7

7 APPENDIX A WING MEASUREMENT VALIDATION...............................................................................91 Objective...................................................................................................................... ...........91 Methods........................................................................................................................ ..........91 Results........................................................................................................................ .............91 Conclusion..................................................................................................................... .........91 B HEMATIN STANDARD CURVE............................................................................................94 Introduction................................................................................................................... ..........94 Objective...................................................................................................................... ...........94 Methods........................................................................................................................ ..........94 Results........................................................................................................................ .............95 Conclusion..................................................................................................................... .........95 C MOSQUITO WINGLENGTH VS. WEIGHT REGRESSIONS...............................................98 Introduction................................................................................................................... ..........98 Objective...................................................................................................................... ...........98 Materials and Methods.......................................................................................................... .98 Conclusion..................................................................................................................... .........99 LIST OF REFERENCES.............................................................................................................102 BIOGRAPHICAL SKETCH.......................................................................................................110

PAGE 8

8 LIST OF TABLES Table page 2-1 Summary statistics for bloodmeal si ze, body size and fecundity of wild Culex nigripalpus used in analyses..............................................................................................30 2-2 Summary of four regressions testing relationships between hematin excreted, winglength and fecundity. Significance values are for the t-statistic...............................31 2-3 Decomposition of effects in path analysis. Note that th e direct effect of bloodmeal size is the portion of its contribu tion not attributable to body size....................................34 3-1 Summary statistics for data measured in regression analyses...........................................42 3-2 Summary of four regressions testing relationships between hematin excreted, winglength and fecundity. ................................................................................................42 4-1 Summary statistics for controlled and measured parameters used in regression analyses....................................................................................................................... .......58 4-2 Summary of 5 regression cal culations of various factors and combinations of factors on fecundity................................................................................................................... ....59 4-3 Summary of two ANCOVA analyses describing slope of the fecundity versus standardized hematin curve in 5 age classes......................................................................64 5-1 Summary statistics for factors and re sponses used in regression models..........................76 5-2 Summaries of five separate regres sion equations predicting fecundity of Culex quinquefasciatus fecundity from individual factors and combinations of factors.............77 5-3 ANCOVA analysis describi ng slope of the fecundity ve rsus winglength relationship for 6 Culex quinquefasciatus age classes...........................................................................81 A-1 Results of linear regression analysis of callipered length against length measured in SigmaScan...................................................................................................................... ....93 B-1. Results of the regression analysis for absorbance versus concentration. The slope estimate was used to calculate all hematin quantities in this work....................................97 C-1. Results of two regression analyses pred icting female mass from winglength. All regressions were significant at the =0.05 level................................................................99

PAGE 9

9 LIST OF FIGURES Figure page 1-1. A. Culex nigripalpus Theobald. B. Culex quinquefasciatus Say........................................21 1-2. Egg raft of Culex quinquefasciatus These eggs are freshly laid, and have not darkened....................................................................................................................... ......22 2-1. Example of a wing photograph used to measure winglengths of female Culex nigripalpus ......................................................................................................................30 2-2. Scatterplot of fecundity vers us unstandardized hematin for wild Culex nigripalpus The line represents the leas t-squares linear regression......................................................32 2-3. Scatterplot of fecundity versus unstandardized winglength for wild Culex nigripalpus The line represents the leas t-squares linear regression......................................................33 2-4. Path diagram describing direct eff ects between the bloodmeal size, body size and fecundity...................................................................................................................... ......34 3-1. Scatterplot of fecundity versus bloodmeal size for Culex quinquefasciatus Data are presented in unstandardized format. Least squares regression line shown.......................43 3-2. Scatterplot of fecundity versus winglength for Culex quinquefasciatus Data are presented in unstandardized format. Least squares regression line shown........................44 4-1. Scatterplot with least squares regressi on line of the effect of age on fecundity of Culex nigripalpus .........................................................................................................................60 4-2. Scatterplot with least squares regressi on line of the effect of body size on fecundity of Culex nigripalpus ...............................................................................................................61 4-3. Scatterplot with least squares regre ssion line of the effect of bloodmeal size on fecundity of Culex nigripalpus ..........................................................................................62 4-4. Scatterplot with least squares regre ssion lines of the effect of bloodmeal size on fecundity of Culex nigripalpus ..........................................................................................63 5-1. Scatterplot of fecundity versus un standardized age at bloodfeeding for Culex quinquefasciatus............................................................................................................... ..78 5-2. Scatterplot of fecundity vers us unstandardized bloodmeal size for Culex quinquefasciatus ................................................................................................................79 5-3. Scatterplot of fecundity versus unst andardized body size for Culex quinquefasciatus........80 5-4. Scatterplot of unstandardized winglen gth versus fecundity showing the response between ages 513 days (solid line, circles) and 17-25 days (broke n line, triangles).......81

PAGE 10

10 5-5. Scatterplot of Percentage hatch versus age for Culex quinquefasciatus ..............................82 A-1. Regression plot showi ng relationship between caliper measurement and SigmaScan measurement.................................................................................................................... ..92 B-1. Example of a spectral scan of Culex quinquefasciatus excrement. Note peak corresponding to maximal absorbance for hematin at 387nm...........................................96 B-2. Standard curve used for calculating am ount of hematin in mosquito excreta. R2 for this regression was 0.9969, F=12830 on 1 and 40 degrees of freedom. N=42.........................97 C-1. Linear regression of wingle ngth on wet mass of pupae of female Culex quinquefasciatus n=42...................................................................................................100 C-2. Linear regression of wingle ngth on wet mass of pupae of female Culex nigripalpus N=43........................................................................................................................... .....101

PAGE 11

11 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 SENESCENCE AND OTHER FACTORS AFFE CT FECUNDITY IN TWO SPECIES OF Culex MOSQUITOES (DIPTERA: CULICIDAE) By Sean Michael McCann December 2006 Chair: Cynthia C. Lord Major Department: Entomology and Nematology Mosquitoes of the genus Culex are considered important vector s of arboviral diseases in the State of Florida. I inves tigated the effects of age, body size and bloodmeal size on the fecundity of Culex nigripalpus and Culex quinquefasciatus using multiple regression models. I found that in both species, body size and bloodmeal size were strong predictors of fecundity in uniform age populati ons, accounting for a large amount of the variance in fecundity observed. I also quantified the direct and indi rect effects of body size on fecundity of wild Culex nigripalpus using a path analysis. This an alysis revealed that there is a strong indirect effect of body size on fecundity mediated by bloodmeal size. In experiments incorporating the effect of age, the responses changed somewhat. In Culex nigripalpus there was an interacti on between age and bloodmeal size predicting fecundity, indicating that the respon se to bloodmeal is not uniform across different aged groups, but rather declines with age. Overall, the number of e ggs produced by this spec ies was found to decline with age. This is the first demonstrati on of an age-related fecundity decline in Culex nigripalpus In Culex quinquefasciatus the response of fecundity to body size is strong and positive in younger age groups, declining as the mosquitoes age. Fecundity overall al so declined with age in this species.

PAGE 12

12 These results demonstrate that fecundity dec lines with age, and in creases with bloodmeal size and body size. Responses to bloodmeal si ze and body size are modified by age. These results can be incorporated into population growth models of these two species, which may aid in better predicting risk of arboviral outbreaks.

PAGE 13

13 CHAPTER 1 INTRODUCTION Mosquitoes Taxonomy The genus Culex is a sizable genus of the Culicidae represented in North America by 29 species. Florida is home to 15 species in this genu s. The two species considered in this thesis are Culex nigripalpus Theobald and Culex quinquefasciatus Say. These fall in the subgenus Culex Linnaeus (Darsie and Ward 2004). The taxonomic status of Culex nigripalpus is uncontroversial, having been described by Fred V. Theobald in 1901 (Knight and P ugh 1973), and undergoing no taxonomic revision since then, except for addressing synonymy. Culex quinquefasciatus on the other hand, has had quite a co lorful history, and even today its specific status is in doubt. At various times it has been considered a subspecies or geographic variant of Culex pipiens Linnaeus, a species, or some co mbination thereof. For a thorough treatment on the history of these taxa, see Vinogr adova (Vinogradova 2000). For the purposes of this paper, Culex quinquefasciatus will refer to Culex quinquefasciatus in the sense of Say 1823. Distribution and Notes on Natural History Culex nigripalpus (Fig. 1-1 A) is an abundant mos quito in the Southeastern USA and ranges into Mexico, the Caribbean Basin and into Central and South America. Culex quinquefasciatus (Fig 1-1 B) is also abundant in the S outheastern USA, but also ranges westward into California. Outside the US, the range could be considered worldwide in warmer regions of the globe. Both are warm weather species, and are found only in warm temperate, subtropical and tropical climates. They are anautogenous species requiring a bloodmeal to provision their eggs.

PAGE 14

14 Eggs are laid together on the surface of the wa ter in batches referred to as rafts, which have a characteristic structure (Figure 1-2). In general the egg rafts of both species are broadly oval, range from 2-4mm in length, and are between 1-1.8 mm in width, with 4-5 rows of eggs (Chadee and Haeger1986). In general, all of the eggs matured in each gonotrophic cycle are laid in one raft, although some may be retained a nd resorbed (Nayar and Knight 1981). In terms of oviposit ion site selection, Culex quinquefasciatus prefers standing water bodies high in organic pollutants, whereas Culex nigripalpus could best be described as a floodwater mosquito, preferring freshly fl ooded grasslands and agricultural areas (Nayar 1982). Both species will also oviposit in othe r types of habitats, especially containers. It is unknown what drives habitat preferences in these species, alth ough olfactory cues are li kely to be important (McCall and Eaton 2001, Olagbemiro et al. 2004). Vector Associations Mosquitoes of the genus Culex are well-known vectors of arb oviral and other diseases in humans and animals (Dow et al. 1964, Nayar 1982, van Riper III et al. 1986, Nayar et al. 1998, Ahid et al. 2000, Turell et al. 2001). Th e first instance of incrimination of a Culex species as a vector of an arboviral disease of humans was in 1933, during an outbreak of viral encephalitis in the city of St. Louis, Missouri (Mit chell et al. 1980). Since that time Culex species have been recognized as important vectors of many serious human ailments such as Japanese Encephalitis in Asia (Rosen 1986), West Nile Virus (WNV) in Africa, Eurasia and the Americas (Sardelis et al. 2001) and St. Louis Encephalitis (SLEV) in the Americas (Day 2001). The species under consideration in this work ar e demonstrated vectors of SLEV and WNV. Culex nigripalpus in particular is generally considered to be the most important vector for SLEV and WNV in the State of Florida (Day 2001, 2005).

PAGE 15

15 Understanding the role a vector plays in tr ansmission cycles of an arbovirus or other disease demands not only an understanding of vect or competence, but also an understanding of the reproductive rates of the vectors. Viral transmission intensity is related to size of the vector population in many arboviral transmission system s (Mitchell et al. 1980, Day 2001), and thus understanding factors that govern population size and increase is important to understanding the viral cycles. Age Age of insect vectors of disease is genera lly investigated because of the relationship between age of vector populations and the proba bility of disease transmission (Mitchell 1983). Most vector-borne diseases requ ire a period of time following expos ure of the vector to reach a stage where the vector is capable of transmission. This is termed the extrinsic incubation period (Meyer 1989), and involves acquisition of the pathogen, propagation a nd development in the vector, and spread of the pathoge n into tissues that allow it to be transmitted (e.g., salivary glands). The likelihood of transm ission of a disease is then influenced by the proportion of the vector population that may have been exposed to the disease and survived the extrinsic incubation period. If the population is older, it is more likely that there are substantial numbers of mosquitoes who have been exposed to the ag ent and survived through the extrinsic incubation period. Aside from vector potential of a population, age has other po pulation-level and individuallevel effects. In many, if not most mosquito species investigated to date, age has been demonstrated to have a negative impact on f ecundity (Jalil 1974, Walte r and Hacker 1974, Akoh et al. 1992, Ferguson et al. 2003, Ma hmood et al. 2004). These studies have investigated age related effects on the fecundity of Culex quinquefasciatus Culex tarsalis Aedes triseriatus and Anopheles stephensi

PAGE 16

16 The processes by which age negatively affects lif e history characteristics such as fecundity are collectively referred to as senescence. Senescence is a phe nomenon that is common to most organisms (Kirkwood and Rose 1991), and is an umbrella term that encompasses many physiological, genetic, and behavioral change s. The question as to why organisms undergo senescence has been described in evolutionary terms (Gavrilov and Gavr ilova 2002). A common explanation is the antagonistic pleiotropy hypothesis. The an tagonistic pleiotropy hypothesis states that genetic mechanisms that enable high ear ly-life fecundity or surv ival may also have the result of lowering late life fecundity or surviv al. These effects would be especially pronounced in organisms with a high level of per diem ex trinsic mortality such as mosquitoes (Dow 1971, Walter and Hacker 1974, Reisen et al. 1991). Thus, one could hypothesize pronounced effects of aging on reproduction in mosquitoes. An alternative to the antagonistic pleiot ropy hypothesis is the mutation accumulation hypothesis (Charlesworth 2000). This theory states that the gradual accumulation of the effects of deleterious mutations over the lifespan of the individual is the reason we observe senescence. Because these have pronounced effects only late in life, there is virtually no selection against them. Note that this theory and the anta gonistic pleiotropy hypothe sis are not mutually exclusive, and in fact predict th e same high rate of senescence in short-lived creatures such as mosquitoes. Others have found that host s eeking and ovipositi on activities of Culex nigripalpus are inhibited during periods of lo w humidity and rainfall (Boi ke 1963, Provost 1969, Day et al. 1989, Day et al. 1990 a). During a drought, mosquitoes seek out sheltered locations to rest. During this time, the age structure of the adu lt mosquito population may become old-biased. Mortality and decreased fecundity in older population may preven t populations of these vectors

PAGE 17

17 from recovering from such drought-induced quiescen ce (inactivity of mos quitoes during periods of low rainfall and humidity) (Shaman et al. 2003) This would have an impact on sustained arboviral transmission by reducing th e abundance of vectors. If, af ter the amplification phase of the transmission cycle there are too few vectors to continue transmission, enzootic maintenance would be unlikely. Bloodmeal Size Several other physiological factors play a role in determining the production of eggs in mosquitoes. Probably the most im portant factor in many species is the size of the bloodmeal obtained. The relevance of this factor is easy to see when one considers that the bloodmeal is the main source of nutrition a female mosquito uses to provision her eggs with yolk. The relationship between bloodmeal size and fecundity is usually strong, and has been established in various studies to follow a linear relations hip (Miura and Takaha shi1972, Edman and Lynn 1975, Akoh et al. 1992, Briegel 2003, Lima et al. 2003, Fernandes and Briegel 2005). The linear relationship between bloodmeal si ze and fecundity has been questioned, as undoubtedly there is an asymptote at the upper end of fecundity representing the maximum number of ovarioles in the mosquito ovary (Miura and Takahashi1972). Another suggested explanation for non-linearity in bloodmeal-fecundity regressions is that the non-linearity represents a diminishing returns function of increasing bloodmeal mass on fecundity gain (Roitberg and Gordon 2005). Roitberg and Gordon f it a constrained, quadratic function to their data, but do not provide an estimate of the improvement in R2 values over a linear model. Bloodmeal Source Another factor that may be important in dete rmining the possible fecundity of mosquitoes is the source of the bloodmeal. Different hosts provide different nutritive values for feeding mosquitoes, and this can determine the number of eggs matured and laid following a bloodmeal

PAGE 18

18 (Mather and DeFoliart 1983). In Culex nigripalpus Nayar and Sauerman (Nayar and Sauerman 1977) showed that egg production va ried with host type, but no ge neralizations can be made. The immune status of a host also may play a role in the expected fecundity return from a given bloodmeal. It has been shown that fecundi ty from bloodmeal sources carrying parasites is less than that from similar, uninfected sources (Ferguson et al. 2003, Lima et al. 2003). In addition, antibodies to mosquito saliva may inhi bit bloodmeal digestion by female mosquitoes (Ramasamy et al. 1988), thereby re ducing the expected fitness gain. Body Size and Teneral Reserves Body size of mosquitoes has been shown to be positively correlated with higher fecundity (Miura and Takahashi1972, Packer and Corb et 1989, Briegel 1990 b, 1990 a, Akoh et al. 1992, Bradshaw and Holzapfel 1992, Blackmore and Lord 1994, Renshaw et al. 1994, Blackmore and Lord 2000, Briegel and Timmermann 2001, Armbrust er and Hutchinson 2002, Telang and Wells 2004). There are several reasons this may be the case. Larger body size is usually the result of greater larval nutrition (Ti mmermann and Briegel 1993, Blackmore and Lord 2000, Briegel 2003), and hence the increase in fecundity with body size indicates mobilization of teneral reserves for first-cycle reproduc tion. Another reason body size a ffects fecundity is that the maximum number of ovarioles is generally greate r in larger insects (Colless and Chellapah 1960, Bonduriansky and Brooks 1999).This means that the larger the body size of an individual mosquito, the more eggs she can mature give n a bloodmeal of optimal size and quality. The effect of body size on maximal number of ovariol es may only be seen in individuals who take the maximum quantity of blood their midguts can allow. In addition, larger mosquitoes can take la rger bloodmeals, and this shows up in many studies as a correlation between body size and bloodmeal size (Hogg et al. 1996, Lima et al.

PAGE 19

19 2003, Fernandes and Briegel 2005). This relationsh ip, when present, can vary in strength (correlation coefficient), but is always of a positive slope. Multiple Factors Several studies have been done that inves tigate the effects of multiple factors on egg production of mosquitoes. In general, multip le regression and MANOVA models improve detection of effects and predictiv e capability because se veral factors are take n into consideration at once. In the literature inve stigating the effects of age along with other factors that affect fecundity, there has been no attemp t to test for interactions be tween factors. It could be hypothesized that the relationship of bloodmeal vol ume to fecundity, or body size to fecundity might change as a result of age, reflecting phys iological changes associat ed with senescence. One example of these might be a change in the e fficiency of bloodmeal digestion or yolk transfer as mosquitoes age, although this has not b een reported. Another mechanism may be the depletion of teneral protein reserves. Insects store protein derived from larval nutrition as hexamerins (Telang et al. 2002), and these stores are generally greater in larger, better fed insects. If storage protein decr eases non-linearly in relation to size with age, it could show up as an interaction between body size and age predicting fecundity. Summary The series of experiments reported in this thesis were designed to determine the contributions of body size, bloodmeal si ze and age to the fecundity of Culex nigripalpus and Culex quinquefasciatus The results of these experiments are of interest from a practical standpoint because the estimates of fecundity re duction with age may be incorporated into models of the population dynamics of these sp ecies, which might improve predictions of outbreaks of arboviral diseases.

PAGE 20

20 They are also of interest from the standpoi nt of improving knowledge about the ecological and physiological parameters gove rning reproductive success.

PAGE 21

21 Figure 1-1. A. Culex nigripalpus Theobald. B. Culex quinquefasciatus Say.

PAGE 22

22 Figure 1-2. Egg raft of Culex quinquefasciatus These eggs are fres hly laid, and have not darkened.

PAGE 23

23 CHAPTER 2 EFFECTS OF BLOODMEAL SIZE AND BO DY SIZE ON FECUNDITY OF WILD Culex nigripalpus Introduction Culex nigripalpus Theobald is a widespread nuisa nce mosquito found throughout the Neotropics. This mosquito is also known to breed in large nu mbers throughout Florida and parts of the Southeastern U.S. It is a capable vector of West Nile Virus (Sardelis et al. 2001) and St. Louis Encephalitis Virus (Day 2001); both arbovirus es are in the family Flaviviridae (Calisher and Karabatsos 1988). Due to its status as a vector of arboviruses, this species of mosquito has been studied intensively for many years (Nayar 1982). Factor s governing its seasonal p opulation cycles have been of great interest, as vector abundance during critical periods of the year can be essential to the continuous transmission of pathogens (Day 2001). Reproductive outp ut in mosquitoes is known to be dependent on several factors such as body size, bloodmeal size, bloodmeal source, infection status and age (Edman and L ynn 1975, Akoh et al. 1992, Lima et al. 2003). Given that resources allocated to reproduction in anautogenous mosquitoes come from the bloodmeal, larval diet and to some degree sugar f eeding in the adult stage (Nayar and Sauerman 1975, Timmermann and Briegel 1999, Briegel 2003), one would predict that these factors would be important predictors of fecundity. This has in fact been demonstrated in many experiments. Few studies have investigated the effect of these factors in wild -caught mosquitoes. In addition, many of these studies have shown that th ere is some dependence between body size and bloodmeal size, yet in the creation of predictive models this depe ndence has not been addressed. If body size is responsible for an increased intake of blood, and th is translates into increased fecundity, then larval nutrition could be regarded as having a greater contribution to lifetime fitness than previously imagined. Data are n eeded to better quantify the resource inputs

PAGE 24

24 important to the maintenance and growth of vector populations, and incorporate these into models of the population dynamics of Culex nigripalpus This study was designed to demonstrate the relationship between body size, bloodmeal size and fecundity, and also to separa te the effects of body size into di rect effects, representing the influence of larval nutrition, and indirect eff ects, representing the influence of body size on bloodmeal size. Materials and Methods Trapping Host seeking females were collected in th e field using two lard can traps hung 1.5 m off the ground at Lockwood Hammock near Vero Beach (27.57572 N, 80.43618 W). This site is well known for producing large numbers of Culex nigripalpus The traps were baited with a live chicken (Production Red strain) placed within the trap in a mesh bag, allowing trapped mosquitoes to bloodfeed (University of Florida IUCUC # D509). The traps were set from 6pm to 8am on June 1, 2006. This timeframe takes acc ount of the evening and morning host-seeking and bloodfeeding habits of this species. Hematin Collection and Analysis Bloodmeal size can be quantified by measuri ng the amount of hematin in the excreta (Briegel 1980). Hematophagous arthropods voi d acid hematin as a byproduct of hemoglobin digestion. It has been found that the quantity of hematin voided correlates in a linear manner with the amount of blood ingested (Briegel 1980, 2003) Therefore, it is an appropriate means of quantifying the relative amount of blood ingested. Other methods including near infrared spectrometry (Hall et al. 1990) and weighing of mosquitoes before and after a bloodmeal (Roitberg and Gordon 2005) were rejected as be ing too time-consuming, invasive, or requiring equipment that was unavailable.

PAGE 25

25 Following trapping, the mosquitoes were br ought to the laborato ry, and 108 bloodfed females were selected and placed into separate 40 ml vials covered with screen. 10% sucrose was provided on small cotton balls, and the females we re given four days to mature eggs in the vial. Temperature was held constant at 27.6 C and relative humidity was 70%. The time required for digestion and egg develo pment by this species has been found to be 72 h at 30C, and 96 h at 24C (Nayar and Knight 1981), and hence four da ys at 27 C was found to be a good balance between maximal survival in the vials and maximal ovarian development. Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00 ml 1% LiCO3. The resulting solutions were d ecanted into spectrophotometric cuvettes and the absorbance at 387 nm was read in a spectrophotometer. Absorban ce readings were converted to micrograms of hematin using a standard cu rve previously prepared (Appendix B). Oviposition After egg maturation, gravid females were tr ansferred into a second set of 40ml vials containing 4.0 ml of 10% (by vol ume) hay infusion in tap water for oviposition. These vials were placed in a screened outdoor enclosure at approximately 5pm. This was done to provide a natural twilight which has been found to induce greater oviposition rates in this species. Oviposition continued for three nights. Each mo rning, egg rafts deposited were removed, placed on water under a microscope and photographed at high magnification with a digital camera. Photographs of the egg rafts were printed out with a standard laserjet pr inter and the number of eggs counted (method suggested by A. Doumboya, personal communication, 2005). Winglengths Following oviposition, females were removed, k illed, identified to species (Darsie 2004), and their wings removed for measurement. Abdom ens were dissected to count retained eggs. The wings were measured by photographing them adjacent to a steel pin of known length

PAGE 26

26 (measured to the nearest thousandth of a m illimeter). The photographs were opened in SigmaScan (SPSS Inc.), scaled using 2-point resc aling, and measured from the alular notch to the distal end of R2, excluding fringe hairs. (Figure 2-1). Analyses All statistical analyses were conducted using S-Plus 7.0 for Windows (Insightful Corp.). Fecundity was scored as the number of eggs in the egg ra ft plus the number of retained eggs, provided the retained eggs numbered fewer than 50. If females retained greater than 50 eggs, they were discarded in the analysis, due to difficulty in counting large numbers of retained eggs (eggs often burst, obscuri ng the slide with opaque yolk). The effect of bloodmeal size (measured by hematin) and body size (measured by winglength) on fecundity was analyzed using simp le linear regressions and a multiple regression of these two factors plus their interaction on fecundity. Non-signi ficant effects and interactions were discarded in a stepwise fashion. Two additional regressions were performed in which the centered, standardized scores (Zscores) of bloodmeal size and body size were used as the predictors (Marquardt 1980), and the unstandardized fecundity as well as the standardized fecundity were used as the responses. To explore the notion that body size has both di rect and indirect e ffects on fecundity, a path analysis was conducted on the fully standardized multiple regression. The hypothesis was that the correlation between body size and bloodm eal size implies that body size has a direct effect on fecundity, and an indir ect effect mediated by bloodmeal size. In this path analysis, body size was considered an exogenous variable and bloodmeal size an endogenous variable. The path analysis presents a causal hypothesis a bout the relationship between the variables, and

PAGE 27

27 then quantifies the effects in the hypothesis. It does not imply that the causal hypothesis is correct (Darlington 1990). Assumptions of homoscedasticity and normality were checked graphically with plots of residuals versus fits and Q-Q plots of the re siduals respectively. A one sample KolmogorovSmirnov Test of Composite Normality was also pe rformed to verify normality of the residuals. Results Only one of the Culex that oviposited in the vials was not Culex nigripalpus This was a female Culex quinquefasciatus and was not used in the analyses. In total, 84 female Culex nigripalpus survived and oviposited in the vials. Of these, 23 retained at least one egg. Overall, egg reten tion was low, averaging just 1.8 eggs/female. Two females retained more than 50 eggs, and these were not included in the an alyses. A summary of data collected on fecundity, winglength and he matin excreted can be found in Table 2-1. Both factors (body size and bloodmeal size) were significant in all re gression analyses at the =0.05 level. Regression equations can be found in Table 2-2. Scatterplots of the unstandardized simple linear regressions are shown in Figures 2-2 and 2-3. What is evident from the standardized regressi on equation is that one standard deviation in winglength is responsible for a gr eater gain in fecundity than a standard deviation in bloodmeal size. With the unstandardized regression, such an interpretation is not intuitive. The unstandardized regression is gi ven because the equation can be used to predict fecundity by inputting normal measures of body size and bloodmeal size. Path Analysis The path analysis indicates that the total contribution of body size to fecundity can be decomposed into a direct effect, and an indir ect effect mediated by bloodmeal size (Table 2-4, Fig. 2-4). This is a quantifica tion of a causal hypothe sis, not a proof of such a hypothesis.

PAGE 28

28 Discussion The mean fecundity for female Culex nigripalpus in this study was quite high (273) compared to the values obtained in 1975 from Tiger Hammock near Vero Beach(Edman and Lynn 1975) (175) and 1977 (Nayar and Sauerman 1977) (160). There are several possibilities for why this difference in fecundity from previous studies was noted: Greater size of bloodmeal: The mosquitoes in this study had access to a host for over 14 h, whereas in the studies cited, the time allowed for feeding was not stated. Strain differences: Perhaps the innate fecundity of the mosquitoes captured in this study was higher than the previous ones cited Strain differences in the chicken: Perhaps th e blood of the chicken used in this study had greater protein content. Unfortunately the strain used in the previous studies was not recorded. Previous bloodfeeding: These mosquitoes, be ing wild caught, may have fed previously and thus had more usable protein to mature eggs than ones raised in the laboratory. This is another instance of bloodmeal size being a strong predictor of fecundity in a mosquito. This is easily understood, as the bl oodmeal is the main source of protein for the development of mature oocytes (Briegel 2003) Edman and Lynn showed that bloodmeal volume is positively correlated with egg maturation in Culex nigripalpus (Edman and Lynn 1975), and many others have shown the same trend in other species of anautogenous mosquitoes (Colless and Chellapah 1960, Cochrane 1972, Miur a and Takahashi1972, Hurd et al. 1995, Lima et al. 2003, Fernandes and Briegel 2005, Roitberg and Gordon 2005). This regression simply provides a more precise accounting of the relati onship between bloodmeal size and fecundity in this species. The multiple regression predicting fecundity fr om bloodmeal size and body size provides a better estimate of potential fecundity than does either factor al one. This reflects the fact that there are two sources of protein for the development of at least the first batch of eggs in any autogenous mosquito: larvally-acquired protein, and protein from the bloodmeal. Taking both

PAGE 29

29 into consideration provides a more precise estimat e of the net contributio n any given female is likely to make to a particular egg batch. The path analysis of the multiple regression indicates a direct contribution of body size and an indirect effect mediated by bl oodmeal size. This would indicat e that body size contributes to fecundity, probably due to the contribution of tene ral reserves. It also contributes by enabling a greater volume of blood, and hence protein, to be ingested during bloodf eeding. This suggests that one of the most important factors determinin g a mosquitos fecundity is the quality of the larval environment. Higher-quality environments (meaning more nutrients available) will result in larger adults with greater reserves (Telang and Wells 2004), ab le to take larger bloodmeals, and producing greater numbers of eggs. This type of relationship should be evident in the wild as Culex nigripalpus varies in size over the year due to ch anges in development time brought about by water temperature cha nges (Day et al.1990 b). Another way that body size might have a posi tive influence on fecund ity is by increasing the maximum number of ovarioles that can be used to produce eggs. It is known that the number of ovarioles in insects is infl uenced by size, and so larger individuals have a higher maximum number of eggs that can be matured each gonotr ophic cycle (Fitt 1990). This would be evident in mosquitoes taking a bloodmeal of maximum si ze and quality, where the great majority of follicles are provisioned with yolk. The total fitness effects of larval habitat qual ity would be greater th an those analyzed in this experiment, as it would affect larval su rvival (Agnew et al. 2000, Reiskind et al. 2004). There is also evidence that la rger adult mosquitoes live longe r (Nasci 1986, Lounibos et al. 1990, Suzuki et al. 1993), and thus size wo uld contribute substantially more to overall fitness than this analysis has shown.

PAGE 30

30 Figure 2-1. Example of a wing photograph us ed to measure winglengths of female Culex nigripalpus Arrow indicates length of measurement, from alular notch (Al) to distal end of wing vein R2. Table 2-1. Summary statistics for bloodmeal size, body size and fecundity of wild Culex nigripalpus used in analyses. Hematin excreted ( g) Winglength (mm) Fecundity Retained eggs Minimum 7.21 2.41 148 0 Mean 18.30 2.95 273 1.83 Maximum 34.57 3.46 385 40 Std. Dev 5.87 0.22 57.13 6.23

PAGE 31

31 Table 2-2. Summary of four regressions testing relationshi ps between hematin excreted, winglength and fecundity. Signi ficance values are for the t-statistic. F-tests for all models were below 0.05. Model Parameter EstimateStd. Error DF T statistic P R2 Intercept 144.93 14.35 10.10 <0.01 Hematin Hematin 7.00 0.75 1,82 9.37 <0.01 0.52 Intercept -310.85 55.19 -5.63 <0.01 Winglength Winglength 197.71 18.63 1,82 1061 <0.01 0.58 Intercept -202.25 50.37 -4.01 <0.01 Hematin 4.14 0.72 5.78 <0.01 Stepwise model*, unstandardized Winglength 135.29 19.12 2,81 7.07 <0.01 0.70 Intercept 273.07 3.44 2,8179.21 <0.01 0.70 Hematin 24.29 4.20 5.78 <0.01 Stepwise model*, regressors standardized, response unstandardized Winglength 29.74 4.20 7.07 <0.01 Intercept 0 0.603 2,810 1.00 0.70 Hematin 0.4251 0.0736 5.78 <0.01 Stepwise model*, all parameters standardized Winglength 0.5205 0.0736 7.07 <0.01 *The interaction between the pred ictors was not significant at the =0.05 level, and hence was removed as part of a stepwise model reduction. Indicates degrees of freedom for overall model F-test

PAGE 32

32 Figure 2-2. Scatterplot of fecundity versus unstandardized hematin for wild Culex nigripalpus The line represents the leas t-squares linear regression.

PAGE 33

33 Figure 2-3. Scatterplot of fecundity versus unstandardized winglength for wild Culex nigripalpus The line represents the l east-squares linear regression.

PAGE 34

34 Figure 2-4. Path diagram descri bing direct effects between the bloodmeal size, body size and fecundity. Table 2-3. Decomposition of effects in path analysis. Note that the direct effect of bloodmeal size is the portion of its contribu tion not attributable to body size. Factor Direct effect Indirect effect Spurious effect Total effect Pearson Correlation coefficient Body size 0.520 0.240 0 0.761 0.761 Effects on Fecundity Bloodmeal Size 0.425 0 0.294 0.425 0.719 Effect on Bloodmeal Body Size 0.565 0 0 0.565 0.565

PAGE 35

35 CHAPTER 3 INFLUENCE OF BLOODMEAL SIZE AN D BODY SIZE ON THE FECUNDITY OF CAPTIVE Culex quinquefasciatus Introduction Culex quinquefasciatus is a ubiquitous peridomestic mosquito in urban and rural environments across the tropics and subtropics. It is a known vector of various filariases (Villavaso and Steelman 1970, Lowrie et al. 1989) as well as viral diseases (Mitchell et al. 1980, Sardelis et al. 2001) and avian malarias (van Ripe r III et al. 1986). Because of its importance in the transmission of pathogens, considerable inte rest has been shown in the bionomics of this species. One measure of the fitness of individual animals is their f ecundity, or reproductive output. Because Culex mosquitoes lay eggs in rafts, usually de positing their entire clutch at once, one can easily estimate the reproductive output of a mo squito by counting the nu mber of eggs in the egg raft. This is the measure is most commonl y used, and relies on the assumption that most mosquitoes in the wild deposit only one clutch (due to high adult mortality). Factors known to affect the fecundity of th is species include bloodmeal size and body size (Lima et al. 2003). Bloodmeals pr ovide the protein as well as some of the carbohydrate and lipid needed to provision a clutch of eggs (Clement s 2000 a). In anautogenous mosquitoes, bloodmeal size is almost always correlated positivel y with fecundity (Cochrane 1972, Miura and Takahashi1972, Edman and Lynn 1975, Mahmood et al. 2004). From a theoretical standpoi nt, body size may correlate with fecundity via several causal mechanisms. The first is that the protein rese rves of larger mosquitoes are generally higher (Briegel 2003). The presence of higher teneral protein reserves means that more protein may be diverted to egg development regardless of the size of the bloodmeal. The second way that body size may be related to fecundity is that larger in sects generally have larger numbers of ovarioles

PAGE 36

36 (Fitt 1990), thus the maximum number of eggs th at can be matured given an optimum bloodmeal is greater. In any case, the si ze of adult female mosquitoes is thought to be principally the result of greater food resources availa ble during larval development. There is also evidence that survival and reproductive success is influenced by size of adult female mosquitoes, so the overall contribution of body size to lifetime fitness may be considerable (Nasci 1986, Lounibos et al. 1990, Suzuki et al. 1993). Experiments have in the past considered bloodmeal size and body size separately, not addressing their combined contribution to fec undity of the mosquitoes. The experiment described here was designed to determine the co mbined effect of these two factors on fecundity of Culex quinquefasciatus using multiple regression analysis. Materials and Methods Larval Rearing Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing approximately 700 ml of tapwater. Pans we re set with 3 egg rafts each of colonized Culex quinquefasciatus from USDA ARS Gainesville FL, esta blished 1995 (Allan et al. 2006). Food was provided daily to each pan as 20 ml of sl urry containing 20 mg/ml 1:1 Brewers yeast/liver powder. This rearing regimen was chosen in an attempt to generate a range of mosquito sizes, while still achieving relatively simultaneous emergence as adults. Pupation and Bloodfeeding Pupae were placed in 500 ml cups containing 100 ml of tapwater, and allowed to emerge in a cage measuring 33 x 33 x 33 cm. Adults were provided with 10% sucrose solution on a cotton wick replaced daily. After 7 days, the mosqu itoes were offered a bloodmeal on a restrained chicken (University of Fl orida IUCUC # D509). Mosquitoes we re allowed to feed to repletion.

PAGE 37

37 Hematin Collection and Quantification Bloodmeal size can be quantified by measuring the amount of hematin in the excreta. Hematophagous arthropods void acid hematin as a byproduct of hemoglobin digestion. It has been found that the quantity of hematin voided corresponds in a linear manner with the amount of blood ingested (Briegel 1980, 200 3). Therefore, it is an appr opriate means of quantifying the relative amount of blood ingested. Other methods including near infrared spectrometry (Hall et al. 1990) and weighing of mosquitoes before and after a bloodmeal (Roitberg and Gordon 2005)were rejected for this study as being too time-consuming, inva sive, or requiring equipment that was unavailable. Immediately following bloodfeeding, individual mosquitoes placed in separate 40 ml vials 2.5 cm diameter, 9.5 cm deep) covered with scre en. 10% sucrose was provided on small cotton balls. Following a 4 day period for egg maturati on, females were transferred to separate vials (see below) for oviposition. The time required di gestion and egg development for this species has been found to be 2-3 days (Elizondo-Quiroga et al. 2006), and hence four days was found to be a good balance between maximal survival in the vials and maximal ovarian development. Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00 ml 1% lithium carbonate. The resulting solutions were decanted into spectrophotometric cuvettes and the absorbance at 387 nm was read in a spectrophotometer. Absorbance readings were converted to micrograms of hema tin using a standard curve prev iously prepared (Appendix B). Oviposition After egg maturation, gravid females were tr ansferred into anothe r set of 40ml vials containing 4.0 ml of 10% hay in fusion (by volume) in tap water for oviposition. The following morning, egg rafts deposited were removed, placed on water under a microscope and photographed at high magnification w ith a digital camera. Photogr aphs of the egg rafts were

PAGE 38

38 printed out using a standard la serjet printer and the number of eggs counted (method suggested by A. Doumboya, personal communication, 2005). Egg rafts were replaced in the vials and in cubated at 27.1C for 36 h and hatched larvae were filtered onto white filter paper and counted. Percentage hatch was calculated as the number of larvae hatched divided by the number of eggs multiplied by 100. Winglengths Following oviposition, females were removed, killed, and their wings excised and mounted on slides for measurement. The wings on the microscope slides were photographed with a standard size reference (a length of steel measured with a calipe r to the nearest thousandth of a millimeter). The photographs were opened in SigmaScan Pro 5 (Systat Software, Inc., Point Richmond, CA), calibrated for size, and measured fr om the alular notch to the distal end of R2, excluding fringe hairs (Packer and Corbet 1989). It was decided to use the di stal end of R2 as a measurement point as it is an unambiguous stan dard landmark. Other studies have used the distal end of the wing, or some kind of other s ubjective measure of the maximal distance (Packer and Corbet 1989, Lima et al. 2003). While others have suggested transforming the winglengths thus obtained by cubing the linear measure (Bri egel 1990 a), the recommendations of Siegel were followed here (Siegel et al. 1992) and winglengths were not transformed. Analyses All statistical analyses were conducted using S-Plus 7.0 for Windows (Insightful Corp.). Summary statistics were calculated for each measured parameter, including minimum, mean, maximum and standard deviation. Fecundity was scored as the number of eggs in the egg raft. The effe ct of bloodmeal size (measured by hematin) and body size (measured by winglength) on fecundity was analyzed

PAGE 39

39 using simple linear regressions and a multiple regression of these two factors plus their interaction on fecundity. If th e interaction was not significant, it was discarded in creation of a predictive multiple regression model. A multiple regression of fecundity on the standardized values (Z-scores) of the predictors was also pe rformed. Z-scores were calculated by subtracting the mean value of a regressor from each observa tion, then dividing this by the standard deviation of the regressor. This produces mean values of zero and standard deviations of one. When used in multiple regression analysis, it allows a more st andard interpretation of slope values, i.e. It apportions mean changes in response due to pr edictor variations of one standard deviation (Marquardt 1980). Doing so allows one to order the predictors in terms of influence on a common scale. Assumptions of homoschedasticity and normality were checked graphically with plots of residuals versus fits and Q-Q plots of the re siduals respectively. A one sample KolmogorovSmirnov Test of Composite Normality was also pe rformed to verify normality of the residuals. A regression analysis testing for linear dependence between the predictors in the multiple regression model was also performed. Results Winglength varied between 2.89 and 3.33mm with a mean of 3.10 and a standard deviation of 0.10 mm (Table 3-1). Hematin voided was al so variable, with a mean of 15.99g and a standard deviation of 4.96 g. Significant simple linear regressions were found for each of the variables (Table 3-2). The predictive power of hematin as a factor predicting fecundity was greater than that for the simple linear regression of winglength pred icting fecundity (Table 3-2). The full multiple regression model predic ting fecundity from both bloodmeal size and body size had a much better fit to the data than either factor consider ed alone. The multiple

PAGE 40

40 regression with the standardized variables as pred ictors indicated that fo r one standard deviation in either variable, the effect of bloodmeal was the more signifi cant source of variation in fecundity. A regression of hematin and winglength on per centage hatch was considered, but rejected due to non-normality of the data and no apparent tr end apparent in scatterplots of the data (data not shown). Discussion The amount of hematin voided by this group of mo squitoes was similar to that in a study conducted on a wild strain from Brazil (Lima et al. 2003). In that study the mean hematin content in excreta varied from 14.60-15.80 g (fed on human blood). Mean fecundity for the Brazilian strain was also much lower, perhaps indicating that the usable protein content for a given quantity of hemoglobin is less in huma n blood than in chicken blood. A study of Culex nigripalpus showed that mean number of eggs per raft was greater with chicken blood than with human blood (Nayar and Sauerman 1977). Alternatively, this could indicate an innate difference in the fecundity of the two strains of mosquito. The slope of the regression of winglength on f ecundity was about twice as great for this strain than for a Brazilian strain fed on hu man blood, although the mean winglength of the Brazilian strain was substantiall y greater than this population (Lima et al. 2003). This may be partially the result of a slightly diffe rent winglength measurement technique. The combined effect of larval nutrition (re presented by adult size) and bloodmeal size (represented by hematin) on fecundity was greater than either of the two f actors alone. In this case, bloodmeal was seen to be a better predictor of fecundity than adult body size, but this may be due to the fact that the mo squitoes raised did not vary gr eatly in body size. A comparison with known body sizes of wild Culex quinquefasciatus was not possible. Ot her studies indicate

PAGE 41

41 that the adult body size in this species is larger (Lima et al. 2003), but these may represent strain differences. It seems likely that the major contributors to female fecundity of this species are bloodmeal volume, bloodmeal source and teneral re serves. This experiment demonstrates that excluding sources of mo rtality, fitness of Culex quinquefasciatus is predicated upon bloodmeal size and teneral size.

PAGE 42

42 Table 3-1. Summary statistics for data measured in regression analyses. Parameter Statistic Hematin (g) Wingle ngth (mm) Fecundity (# of eggs) percentage hatch Min 5.57 2.89 76 7 Mean 15.99 3.10 163 85 Max 30.51 3.33 256 100 N 71 71 71 71 Std. Deviation 4.96 0.10 36 20 Table 3-2. Summary of four regressions testing relationshi ps between hematin excreted, winglength and fecundity. Signi ficance values are for the t-statistic. F-tests for all models were below 0.05. Model Parameter CoefficientStd. Error DF T statistic P R2 Intercept 79.79 10.31 7.74 <0.01 Hematin Hematin 5.20 0.62 1,69 8.44 <0.01 0.51 Intercept -200.19 121.19 -1.65 0.10 Winglength Winglength117.10 39.06 1,69 3.00 <0.01 0.11 Intercept -319.35 78.16 -4.09 <0.01 Hematin 5.319 0.53 10.09 <0.01 Stepwise model*, unstandardized Winglength128.10 24.92 2,68 5.14 <0.01 0.65 Intercept 162.93 2.59 2,6862.86 <0.010.65 Hematin 26.36 2.61 10.09 <0.01 Stepwise model*, regressors standardized, response unstandardized Winglength13.43 2.61 5.14 <0.01 *The interaction between the pred ictors was not significant at the =0.05 level, and hence was removed as part of a stepwise model reduction. Indicates degrees of freedom for overall model F-test.

PAGE 43

43 Figure 3-1. Scatterplot of fec undity versus bloodmeal size for Culex quinquefasciatus Data are presented in unstandardized format. Least squares re gression line shown.

PAGE 44

44 Figure 3-2. Scatterplot of f ecundity versus winglength for Culex quinquefasciatus Data are presented in unstandardized format. Least squares regression line shown.

PAGE 45

45 CHAPTER 4 Culex nigripalpus AGING AND FECUNDITY Introduction Culex nigripalpus is a species of mosquito distribu ted throughout the Neotropics and in large portions of the Caribbean Basin and the So utheastern United States (Nayar 1982). It has been incriminated as a vector of St. Louis En cephalitis Virus (Dow et al. 1964) and West Nile Virus (Sardelis et al. 2001), as well as several ot her viral, protozoan, and helminth diseases of man and animals (Nayar 1982, Nayar et al. 1998). It is believed that this species is the major enzootic vector of both WNV a nd SLE in Florida (Day 2001, Rutle dge et al. 2003). Considering its role in the transmi ssion of arboviruses in Florida and othe r places, much attention has been focused on the bionomics and population dynamics of this species (Dow 1971, Nayar 1982, Day et al. 1990 b, Day 2001). Factors affecting the fecundity (egg production) of mosquitoes have been investigated for many years. In general, larval nutrition, adult nu trition and age are considered important factors (Clements 2000 b, Briegel 2003). A portion of th e protein reserves ac quired during larval development may be used for the provisioning of eggs (Briegel 2003). Larval nutrition is difficult to quantify in the field, but it has b een shown that the size of an adult mosquito correlates well with the amount of nutrients av ailable to the larvae (Akoh et al. 1992, Blackmore and Lord 1994, Agnew et al. 2000, Blackmore and Lord 2000, Armbruster and Hutchinson 2002). Because of this, adult size serves as a convenient proxy for teneral protein reserves. In addition to teneral protein reserves, anaut ogenous mosquitoes also use the bloodmeal as a source for protein used to mature eggs. Bloodm eal size has usually been highly correlated with fecundity in many mosquito species studi ed to date, and this holds true for Culex nigripalpus

PAGE 46

46 (Edman and Lynn 1975, Ferguson et al. 2003). Meas uring both bloodmeal size and body size has the potential to address the contributions of tw o important sources of variation on fecundity. This effect of age on fecundity can be seen as an aspect of senescence, the umbrella term encompassing all deleterious effects of time on fitness of an individual. Age has been demonstrated to have an effect on fecundity in Culex mosquitoes (Walter and Hacker 1974, Suleman and Reisen 1979, Akoh et al. 1992), bu t whether age modifies the response to bloodmeal size or body size has not been investigated. It is conceivable that physiological systems mediating fecundity such as digestion or mobilization of teneral reserves are impacted by senescence. In any case, generalizing the re sponses of species studied so far, fecundity should increase with bloodmeal size and body size, a nd decrease with age of the mosquito. The purpose of this experiment was to analyze the response of fec undity to these three factors (bloodmeal size, body size, and age) combined, as well as any interactions in a linear multiple regression model. Materials and Methods Larval Rearing Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing approximately 700 ml of tapwater. Pans were set with 3 egg rafts each from a colony of Culex nigripalpus The colony was established from a Vero Beach FL collection (Allan et al. 2006) in 1999 (Erin Vrzal, USDA, personal communication). Food was provided daily to each pan as 20 ml of slurry containing 20 mg/ml 1:1 Brewers yeast/liver powder. Pupation Pupae were placed in 500 ml cups in a la rge cage measuring 57 x 57 x 57cm and sugar was provided to the emerging adults as 10% sucrose so lution on cotton wicks, replaced daily. Adults were allowed to emerge for 12 h following the emergence of the first female, whereupon the

PAGE 47

47 cups were removed. The final density of mosqu itoes in the cage was estimated to be about 500 females. Immediately following the removal of the cups, the cage was placed in a separate room with windows, as it has been noted that succe ssful mating in this colony is inhibited if mosquitoes are denied access to either sunlight or twilight (p ersonal observation). Following 2 days in this separate room, the cage was return ed to the controlled-environment chamber where temperature was maintained at 26.0 1.9 C for the duration of the experiment. Humidity was 88.5 7.5% RH. Light cycle was 16:8 L:D. Bloodfeeding Beginning at 5 days post-eclosion, varying numbe rs of host-seeking females were removed from the cage and bloodfed on a restrained chicke n (University of Florid a IUCUC # D509). This was repeated at four day intervals until 6 groups were obtained. The oldest group was bloodfed at 25 days post-eclosion. The numbers remove d for bloodfeeding were ad justed upwards with age to anticipate higher mortality of older indi viduals. Mortality was severe in the oldest age group, with only 1 out of 66 bloodfed females surv iving to oviposit, compared with 30 of 31 females surviving to ovipos it in the youngest age group. Termination of Study The study was terminated when the combin ation of mortality and removal of hostseeking females depleted the supply in the larg e cage. As a result, ages from 5-25 days posteclosion were tested. Age at ovi position was calculated by simply a dding four days to the age at bloodfeeding. Hematin Collection and Quantification Bloodmeal size was quantified by measuring th e amount of hematin in the excreta. Hematophagous arthropods void acid hematin as a byproduct of hemoglobin digestion. It has been reported that the quanti ty of hematin voided by many mosquitoes corresponds in a linear

PAGE 48

48 manner with the amount of blood ingested (Briegel 1980, 2003). Therefore, hematin analysis is an appropriate means of quan tifying the relative amount of blood ingested. Other methods including near infrared spectrometry (Hall et al 1990) or weighing of mosquitoes before and after a bloodmeal (Roitberg and Gordon 2005) were rejected for this study as being too timeconsuming, invasive, or requiring equipment that was unavailable. Immediately following bloodfeeding, i ndividual mosquitoes were plac ed in separate 40 ml vials 2.5 cm diameter, 9.5 cm deep) covered with scre en. 10% sucrose solution was provided on small cotton balls, replaced daily. Following a four day period for egg maturation, females were transferred to separate vials (s ee below) for oviposition. The time required for digestion and egg development for this species has been found to be 72 h at 30C, and 96 h at 24C (Provost 1969, Nayar and Knight 1981), and hence four days at 27 C was found to be a good balance between maximal survival in the vials a nd maximal ovarian development. Fecal material in the vials where egg maturati on took place was rinsed from the vial with 2.00 ml 1% LiCO3. The resulting solutions were decanted in to spectrophotometric cuvettes and the absorbance at 387 nm was read using a spectrop hotometer. Absorbance readings were converted to micrograms of hematin using a standard curve previously prep ared (Appendix B). Oviposition After egg maturation, gravid females were tr ansferred into a second set of 40ml vials containing 4.0 ml of 10% (by vol ume) hay infusion in tap water for oviposition. These vials were placed in a screened outdoor enclosure at approximately 5pm. This was done to provide a natural twilight which has been found to induce gr eater oviposition rates in this species (personal observation). If females did not oviposit on the firs t night, they were allowed a second night to oviposit. The following morning, egg rafts de posited were removed, placed on water under a microscope and photographed at hi gh magnification with a digital camera. Photographs of the

PAGE 49

49 egg rafts were printed out using a standard la serjet printer and the number of eggs counted (method suggested by A. Doum boya, personal communication, 2005). Egg rafts were replaced in the vials in which they were laid and incubated at 27.1C for 36 h to allow hatching (Provost 1969) Hatched larvae were filtere d onto white filter paper and counted. Percentage hatch was calculated as the number of larvae hatched divided by the number of eggs multiplied by 100. Winglengths Wings were photographed with a digital camera adjacent to a steel pi n of known size. The wing photographs were opened in SigmaScan Pr o 5 (Systat Software, Inc., Point Richmond, CA), calibrated for size using a 2-point rescaling function, and measured from the alular notch to the distal end of R2, excluding fringe hairs (Packer and Corb et 1989). It was decided to use the distal end of R2 as a measurement point rather than the wingtip, as it is an unambiguous standard landmark. While others have suggested transf orming the winglengths thus obtained by cubing the linear measure (Briegel 1990 a), the recomme ndations of Siegel were followed here (Siegel et al. 1992), thus winglengt hs were not transformed. Analyses All regression analyses we re conducted using S-Plus 7.0 for Windows (Insightful Corp.). ANCOVA slope estimates were tested in SAS version 9.00 for Windows. Fecundity was scored as the number of eggs in the egg raft plus the number of retain ed eggs in the killed females, provided these numbered less than 50. If females retained more than 50 eggs, they were discarded in the analysis, due to difficulty in counting large numb ers of eggs (eggs often burst, obscuring the slide with opaque yolk). Simple linear regressions of each predictor on fecundity were performed to compare with published results of these predictors on fecundity.

PAGE 50

50 A multiple linear regression was performed on st andardized variables regressing fecundity on body size, bloodmeal size and age (Marquardt 1980) Non-significant interactions and main effects (if appropriate) were di scarded in a stepwise fashion. Z-scores were calculated by subtracting the mean value of a regressor fr om each observation, then dividing this by the standard deviation of the regressor. This produc es mean values of zero and standard deviations of one. When used in multiple regression analysis, it allows a more standard interpretation of slope values (Marquardt 1980), i.e. It apportions mean changes in response due to predictor variations of one standard devi ation. It also allows for more adequate assessment of model quality and provides better insi ght for stepwise model reducti on. For comparative purposes, a multiple regression of fecundity on the untra nsformed variables was also performed. Normality was confirmed with a KolmogorovSmirnov Test of Composite Normality of the residuals, and homoschedasticity was verified vi sually with a plot of residuals versus fitted values. Significant interactions with age were explored by examining scatterplots of the interacting variables, and generating slope estimates by ANCOVA. The ANCOVA was set up using the recommendations of Huitema (Huitema 1980), and wa s done to illustrate th e differences in slope of the simple regression of one va riable predicting fecundity at di fferent levels of the interacting variable. The mosquitoes aged 21 days and 25 days were grouped toge ther, because only one mosquito oviposited at 25 days, and thus th ere was no way to generate a slope estimate. Results The winglengths of female Culex nigripalpus in this experiment ranged from 2.614mm to 3.227mm (Table 4-1). The range wa s less than that of a wild cohor t captured in a lard can trap (Chapter 2, Table 2-1), but the means were simila r. The amount of hematin was also variable,

PAGE 51

51 but again was less so than the mosquitoes captured in the study of wild Culex nigripalpus Fecundity ranged from 47-262 eggs with a mean of 171. Each regression tested was significant at the =0.05 level (Table 4-2). Simple linear regressions of fecundity against hematin excreted, winglength and age were all significant, with directions as predicted by theor y. The slope of the simple linear regression of fecundity versus age was -4.37, meaning that with every day of ag e added, mean fecundity would be expected to decline by 4.37 eggs. This simple linear regression had the highest R2 of any of the three performed, yet even this was low compared to the multiple model. In the multiple regression models, there wa s a significant interaction between age and hematin after stepwise model reduction (Table 42). The standardized full model provides the clearest picture of the relationships between the parameters, showing a negative correlation between age and fecundity (Figure 4-1A), and positive correlations between fecundity and both winglength and hematin (Figures 4.1B, C). On the other hand, the unstandardized model shows a somewhat confusing result, na mely that age has a positive sl ope, and is not a significant predictor. This is largely due to the fact th at models with higher-order terms (including interactions) suffer from scales that have orig ins far from the centroid of the observed data (Marquardt 1980). In the case of the unstandardi zed multiple regression model, each coefficient is calculated holding every other fact or at zero, which is not a value attained in this dataset. In the case of the standardized model, each coeffi cient is calculated at the mean of each other variable, providing a truer approximation of mean effect. For both qual itative and quantitative interpretation, the standardized regression is preferable. The uns tandardized form may be used as a simple predictive equation.

PAGE 52

52 Further analysis of the signi ficant interaction by contrasti ng the slopes of the fecundity versus hematin curves of the different age classes reveals a complex picture, where slopes for the hematin fecundity regression were significantly di fferent from zero at 5, 9, and 17 days, but not at 13 days or the pooled values for 21-25 days (Table 4-3). Slope estimates, regardless of significance, were positive until the group of mosquitoes aged 21-25 days (Figure 4-4). Discussion Age in insect vectors of disease is generally st udied in order to appraise the potential of a population to transmit disease. Because viral and other diseases require an extrinsic incubation period in the vector in order fo r that vector to become capable of transmission, aging populations are considered an important component in the di sease transmission cycle. The effect of age on fecundity of a vector population has been less often studied, but one may easily hypothesize significant epidemiological consequences to altera tions in reproductive rates of vectors. In certain circumstances, large numbers of insect ve ctors of disease need to be present over a transmission season in order for major outbrea ks to occur (Mitchell et al. 1980, Day 2001). To date there have been no published studi es on the effect of age on fecundity of Culex nigripalpus The regression coefficient for the simple linear regression of age on fecundity in this study is therefore the first documentation of an age effect on reproduction of this species. The slope of the regression line (-4.374) compar es well with a published study on a Vero Beach strain of Culex quinquefasciatus which gave a slope of -4.146 for the decrease in fecundity due to age (Walter and Hacker 1974). Estimates for th e rate of decrease in fecundity with age vary by species and strain, and thus th e value reported in this thesis is only representative of the particular captive strain used. It may be that w ild populations or different isolated captive strains may respond to age in a differe nt manner than reported here.

PAGE 53

53 Culex nigripalpus is hypothesized to have substantia l survival during drought conditions, when oviposition sites are scarce (Day et al. 1990). It is also known that ac tivity patterns of this species are correlated with rainfa ll. If drought conditions prevail in the field, and oviposition and bloodfeeding do not occur, the females that survive the drought will be older at reproduction than if inclement conditions had not occurred. In this scenario, one could expe ct to see lower mean clutch sizes and hence lower popul ation growth in an aging population when compared to a similar sized cohort of younger females. This variability in reproductive output with age may be significant, but it can be argued that the effect of mortality w ith age would far outstrip the redu ction in fecundity. An estimate for the daily rate of survival for this species was made with a recapture study of F1 adults marked with radioactive phosphorus (Dow 1971). The estimate arrive d at in this study was 81%, meaning that the daily mortality rate is 19%. This estimate may be biased due to the fitness consequences brought about by exposure to ra diation, but it serves as a reminder that survivorship of wild Culex can be quite low. Notwithstandi ng the sizable contribution of mortality to overall fitness parameters, the effect of age on fecundity would still be additive to any population fitness reductions brought about by mortality. Size of bloodmeal and size of adult female mosquitoes have often been reported as significant predictors of fecundit y. It has been reported that bloodmeal volume affects the firstcycle fecundity of Culex nigripalpus if one measures this parame ter by visual estimation (Edman and Lynn 1975), but the methods used in that stu dy did not allow calculation of a slope for that relationship. This study first confirms, then ex tends the precision of this finding by providing a means to quantify the amount of blood ingested and to analyze the relationship of this factor with fecundity. Overall, the two simple regressions of winglength and hematin on fecundity were

PAGE 54

54 significant, but when analyzed wi thout considering the effect of age, they had relatively little predictive power (Table 4-2). The full model for prediction of fecundity from age, body size and bloodmeal size had significantly greater predictive power than a ny of the simpler models, and fitting a model including the significant inter action added approximately 4% to the predictive ability of the equation (Data not shown). Before considering the interaction al one, it is important to first discuss the relative merits of the two full mode ls developed. In the uns tandardized model, the predictors are scaled in the familiar units of millimeters of winglength and micrograms of hematin excreted. It would seem that using this model for analysis would be sufficient, and it is, providing predicting fecundity from th ese factors is all that is requi red. If one wants to give the relationship more consideration, such as what a typical change in ei ther of these parameters does to fecundity, the unstandardized model is no longer sufficient. In this dataset, there are no deviations of a whole millimeter in winglength between any two female Culex nigripalpus The typical difference in winglength between any tw o randomly chosen females is only 0.13 mm (Table 4-1). Using variables c oded as Z-scores allows one to quickly determine what effect a typical (1 standard deviation) change in a pred ictor will have on the re sponse. Since all the predictors are Z-scores, one can compare between typical variations and estimate the relative importance of each predictor to changes in the res ponse. Failure to standardize predictors also has the unwelcome effect of obscuring significa nt effects of factors by making their partial slopes seem small in comparison to others. This is only an effect of scaling, and disappears when factors are properly sta ndardized (Marquardt 1980). In addition, a regression using Zscores eliminates most of the non-essentia l ill-conditioning (multic ollinearity) between predictors that cause higher order models to behave erratically (Marquardt 1980).

PAGE 55

55 An interaction between bloodmeal size and age has never been reported before in mosquitoes, nor has such a possibi lity been considered. Therefor e, the fact that a significant interaction was found is both surprising and novel. However, the interaction was mainly due to the last age group considered, and due to low su rvival, the power for estimating the effect was low. Consideration of the inte raction by grouping the mosquitoes into two arbitrary categories of young and old demonstrates that in the ol der age groups, bloodmeal size has no significant effect on fecundity, whereas in the younger age groups, bloodmeal si ze was a significant predictor of fecundity. Possible mechanisms underlying this interaction are difficult to hypothesize. Empirically, one can describe the interaction by saying that the reproductive advantage conferred by a larg er bloodmeal is only eviden t in younger mosquitoes. As mosquitoes age, the advantages diminish until th ere is no significant rela tionship between size of bloodmeal and fecundity. This type of interaction may only be evid ent in experimental settings such as these, where mosquitoes are denied a bloodmeal until a given age, or this result may generalize to aging mosquitoes regard less of their access to bloodmeals. A possible physiological mechanism for this ch ange in the relations hip between bloodmeal size and fecundity with age is an overall decline in the ability of an insect to synthesize protein as it ages (Levenrook 1986). Protein enzymes are utilized in bloodmeal digestion (Clements 2000 c) as well as in nutrient tran sport to maturing oocytes (Clement s 2000 a). If the efficiency of these systems is substantially reduced, the ove rall advantage of larger bloodmeals may be negated. A simple failure of the physiological systems underlying bloodm eal digestion and/or nutrient transport would account for the observe d interaction between bloodmeal size and age predicting fecundity. It is intere sting to note that in a study of Anopheles stephensi Ferguson and colleagues (2003) noted that the positive cont ribution of bloodmeal to fecundity disappeared

PAGE 56

56 if the mosquito was infected with Plasmodium chabaudi Perhaps the state of senescence and the state of Plasmodium infection both affect digestion or reserve mobilization. The novel finding of an interaction between bloodmeal size and age predicting fecundity should be investigated further. Specifically, a study similar to this one, but utilizing a greater number of age classes might better pinpoint the ag e at which increased bloodmeal size fails to result in increased fecundity. Another area that would be interesting to pur sue would be to design an experiment where multiple bloodmeals and ovipositions are provided and recorded. This would be a more complex experiment, but it would allow for the quantificat ion of the effect of age on total lifetime fecundity in an experimental setting that is si milar to a situation encountered by wild populations of mosquitoes. It is unlikely that wild mosquito es living to an age of 17 days or more would not have had the opportunity to bloodfeed. A llowing for early and frequent bloodfeeding opportunities in an experimental setting would he lp determine whether protein input from early bloodmeals would better preserve digestive efficien cy in later bloodmeals. It is at present unclear whether protein resource s acquired from bloodmeals can be used by a mosquito for its own cellular maintenance and synthesis of enzyme s. Mosquitoes can acquire extra-ovarian lipid and carbohydrate from bloodmeals, but studies on u tilization of bloodmeal protein suggest that diversion of this resource to maternal reserves is minimal in Anophelines and non-existent in Culicines (Clements 2000 b). Conclusions Good prediction of firs t-cycle fecundity of Culex nigripalpus may be achieved by using bloodmeal size, body size and age as predictors, within the age range of 5 to 25 days of age. It is likely that these factors are important in determini ng fecundity in wild popul ations of this species as well. The contribution of bloodmeal size to fe cundity declines with ag e of this mosquito.

PAGE 57

57 This is a novel finding that bears closer scruti ny to determine if it hol ds across different agerelated nutritional regimes.

PAGE 58

58 Table 4-1. Summary statistics fo r controlled and measured pa rameters used in regression analyses. Parameter Statistic WinglengthAgeHematin Excreted Fecundity Minimum 2.6151.12 47 Mean 2.8611.811.63 171 Maximum 3.232519.25 262 N 666666 66 Standard Deviation 0.134.963.98 45

PAGE 59

59 Table 4-2. Summary of 5 regressi on calculations of various factor s and combinations of factors on fecundity. Model Parameter CoefficientStd. Error DF T or F P R2 Intercept 222.48 12.72 17.4951 <0.01 Age -4.37 0.9955 -4.3937 <0.01 Model 39.81 1,6419.3 <0.01 Age 0.23 Intercept -96.03 118.5271 1,64-0.8102 0.42 Winglength 93.27 41.3709 2.2546 0.03 Model 43.71 5.083 0.03 Winglength 0.07 Intercept 122.86 16.1854 1,647.5906 <0.01 Hematin 4.13 1.3176 3.1364 <0.01 Model 9.837 <0.01 Hematin 0.13 Intercept -110.35 97.3730 -1.1333 0.2615 Winglength 73.45 32.1597 2.2839 0.026 Hematin 11.15 2.9717 3.7529 <0.01 Age 1.86 2.6099 0.7147 0.48 Hematin*Age-0.58 0.2277 -2.5671 0.01 Model 33.57 4,61 <0.01 Stepwise model, unstandardized 0.48 Intercept 171.43 4.1363 41.4451 <0.01 Winglength 9.63 4.2148 2.2839 0.03 Hematin 16.97 4.2325 4.0085 <0.01 Age -24.47 4.2610 -5.7423 <0.01 Hematin*Age-11.54 4.4951 -2.5671 0.01 Stepwise model, regressors standardized Model 4,61 <0.01 0.48 *Effects not significant at the =0.05 level discarded Indicates degrees of freedom for overall model F-test.

PAGE 60

60 510152025Age (Days) 0 50 100 150 200 250Fecundity (Eggs laid and retained) Fecundity=222.48 4.37 (age) Figure 4-1. Scatterplot with least squares regres sion line of the effect of age on fecundity of Culex nigripalpus

PAGE 61

61 2.52.62.72.82.93.03.13.2Winglength (mm) 0 50 100 150 200 250Fecundity (Eggs laid and retained) Fecundity=-96.03 + 93.27(Winglength) Figure 4-2. Scatterplot with leas t squares regression line of the e ffect of body size on fecundity of Culex nigripalpus

PAGE 62

62 381318Hematin ( g) 0 50 100 150 200 250Fecundity (Eggs laid and retained) Fecundity=122.86+ 4.13 (Hematin) Figure 4-3. Scatterplot with least squares regression line of the effect of bloodmeal size on fecundity of Culex nigripalpus

PAGE 63

63 Figure 4-4. Scatterplot with least squares regression lines of the effect of bloodmeal size on fecundity of Culex nigripalpus Note that in the ages 2125 days, slope estimate was negative, although the regression was not significant at the =0.05 level.

PAGE 64

64 Table 4-3. Summary of two ANC OVA analyses describing slope of the fecundity versus standardized hematin curve in 5 age classe s (slope at 25 days was not estimated due to low sample size). Also shown are the slopes of the pooled young versus old mosquitoes. Note that the slope in the old group is not signifi cantly different from zero. Age (Days) Slope EstimateStd. ErrornT P 5 6.0022.6801132.24 0.0291 9 8.4251.951174.32 <0.01 13 1.6721.766190.95 0.3479 17 5.2592.156122.44 0.0179 21-25 -7.9584.3735-1.83 0.0732 Ages 5-13 4.421.41493.13 <0.01 Ages 17-25 2.892.36171.23 0.2250

PAGE 65

65 CHAPTER 5 SENESCENCE AND FECUNDITY OF Culex quinquefasciatus Introduction Culex quinquefasciatus Say is a cosmopolitan mosquito species found in the tropics and subtropics. It is a noted vector of diseases, includi ng filariases (Ahid et al. 2000, Lima et al. 2003), protozoan parasites (van Ri per III et al. 1986), and various arboviruses (Meyer et al. 1983, Sardelis et al. 2001). Due to its status as a vect or, much attention has been paid to the bionomics of this species. Population density and dynamics of vectors such as this are of interest because they are one element affecting the probability of disease tran smission (Day 2005). Many factors are known to affect the growth a nd maintenance of populations of Culex quinquefasciatus at the level of landscapes, microhab itats, and the physiology of indi vidual mosquitoes. This study focuses on some of the important predictors of reproductive output at the level of adult physiology, as the ultimate determinant of populati on success is success of individuals (Briegel 2003). Like all anautogenous mosquitoes, Culex quinquefasciatus depends on a bloodmeal for the necessary nutrients (protein) to produce eggs. There have be en many studies detailing the relationships between bloodf eeding and reproductive output (Akoh et al. 1992, Hogg et al. 1996, Roitberg and Gordon 2005). Studies on Culex quinquefasciatus have documented a positive relationship between bloodmeal size and fecundity (Akoh et al. 1992, Lima et al. 2003). This indicates that the size of the bl oodmeal, in part, determines the num ber of eggs that can be laid. Other factors are known to influence mosqu ito fecundity, including body size and teneral reserves of the adult females (Briegel 2003). A portion of the protein reserves accumulated during the larval period can be used for nourishm ent of eggs, and so better larval conditions usually result in female mosquitoes capable of greater reproductive output Body size is usually

PAGE 66

66 positively correlated with larval nut rition, so it serves as an indicat or of larval habitat quality, and hence teneral reserv es (Briegel 2003, Telang and Wells 2 004). Many studies have shown a positive relationship between female body size and egg production (Briegel 1990 a, 1990 b, Akoh et al. 1992, Lyimo and Takken 1993, Armbrust er and Hutchinson 2002, Lima et al. 2003). Other members of this species complex, such as Culex pipiens molestus Forskal, a species in the same complex of species (Vinogradova 2000), derive the entire protein input for egg production from larval nutrition (a re autogenous). It is not unr easonable to assume that a portion of the protein required for egg production in the anautogneous members of this complex, such as Culex quinquefasciatus is also derived from the larval stage. A factor which is generally considered to lowe r reproductive output in many animal taxa is age. In animals with high rates of daily mortality such as Culex mosquitoes (Dow 1971, Elizondo-Quiroga et al. 2006), selection is likely to have favored high early-life reproduction, at the expense of late-lif e fitness and reproductive capacity (Kirkwood and Rose 1991). This follows from a theory known as the antagonist ic pleiotropy hypothesis (Williams 1957). This hypothesis states that mechanisms favoring early life fecundity ma y have deleterious effects in later life. This is especially pronounced in animals with high per di em mortality such as mosquitoes, since the numbers surviving to a late age are insignificant, as is the contribution of these individuals to population growth. So far the physiological mechanisms underlying fecundity reductions with age have not been determined. A number of studies have examined declines in reproduction with age in mosquitoes, and several of these (Walter and Hacker 1974, Gomez et al. 1977, Suleman and Reisen 1979, Akoh et al. 1992) have detailed such d eclines in various strains of Culex quinquefasciatus Most of the studies just cited have not examined the role of other physiological parameters such as body size

PAGE 67

67 and bloodmeal size, or have done so in a manne r that does not include these other relevant factors in a global model. Because of this, it is unclear what the relative contributions of each of these factors to fecundity in an aging population tr uly are. The implicatio n of this decline in reproductive capacity with age for the population grow th rate is unclear, but it may at times be severe (Charlesworth 2000). In the interest of developing predictive models for fecundity declines with age, consideration of other factors such as bloodmeal size and body size may improve the robustness and predictive qualities of the model. This study was designed to determine the contributions of age, body size and bloodmeal size to the fecundity of Culex quinquefasciatus considered together in a multiple regression context. Materials and Methods Larval Rearing Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing approximately 700 ml of tapwater. Pans we re set with 3 egg rafts each of colonized Culex quinquefasciatus from USDA ARS Gainesville FL, esta blished 1995 (Allan et al. 2006). Food was provided daily to each pan as 20 ml of sl urry containing 20 mg/ml 1:1 Brewers yeast/liver powder. This rearing regimen was chosen in an a ttempt to generate a range of sizes, while still achieving relatively simultaneous emergence as adults. Pupation Pupae were placed in 500 ml cups in a la rge cage measuring 57 x 57 x 57cm and sugar was provided to the emerging adults as 10% sucrose solution on cotton wicks. Adults were allowed to emerge for 12 h following the emergence of the first female, whereupon the cups were removed. The final density of mosquitoes in th e cage was estimated to be about 700 females.

PAGE 68

68 Temperature was 26.8 C for the duration of the experiment. Relative humidity was 91.7%, and the light cycle was 16:8 L:D. Bloodfeeding Beginning at 5 days post-eclosion, approximate ly 36 host-seeking females were removed from the cage and bloodfed to repletion on a rest rained chicken (Univers ity of Florida IUCUC # D509). This was repeated at 4 day intervals until 6 groups were obtained. The age range thus produced was 5-25 days post-eclosion. Hematin Collection and Quantification Bloodmeal size can be quantified by measuring the amount of hematin in the excreta. Hematophagous arthropods void acid hematin as a byproduct of hemoglobin digestion. It has been found that the quantity of hematin voided corresponds in a linear manner with the amount of blood ingested (Briegel 1980, 2003). Determina tion of the amount of hematin in the excreta thus provides an estimate of the relative amount of blood ingested. This method is easy to implement on large numbers of mosquitoes and is minimally invasive. Immediately following bloodfeeding, individual mosquitoes placed in separate 40 ml vials (2.5 cm diameter, 9.5 cm deep) covered with sc reen. 10% sucrose was provided on small cotton balls. Following a 4 day period for egg maturati on, females were transferred to separate vials (see below) for oviposition. The time required di gestion and egg development for this species has been found to be 2-3 days (Elizondo-Quiroga et al. 2006), and hence four days was found to be a good balance between maximal survival in the vials and maximal ovarian development. Fecal material in the vials where egg maturation took place was rinsed from the vial with 2.00 ml 1% LiCO3. The resulting solutions were d ecanted into spectrophotometric cuvettes and the absorbance at 387 nm was read us ing a spectrophotometer. Absorb ance readings were converted to micrograms of hematin using a standard curve previously prep ared (Appendix B).

PAGE 69

69 Oviposition After egg maturation, gravid females were tr ansferred into a second set of 40ml vials containing 4.0 ml of 10% (by vol ume) hay infusion in tap water for oviposition. The following morning, egg rafts deposited were removed, placed on water under a microscope and photographed at high magnification w ith a digital camera. Photogr aphs of the egg rafts were printed out using a standard la serjet printer and the number of eggs counted (method suggested by A. Doumboya, personal communication, 2005). Egg rafts were replaced in the vials and inc ubated at 27.1C for 36 h to allow hatching. Hatched larvae were filtered onto white filter paper and counted. Percentage hatch was calculated as the number of larvae hatched divided by the number of eggs multiplied by 100. It is unknown whether the handling of the egg rafts a ffected percentage hatch. Delayed hatching is not known from subgenus Culex mosquitoes, but has been observed in the subgenus Melanoconion (Hair 1968). Winglengths Following oviposition, females were removed, killed, and their wings excised and mounted on slides for measurement. Abdomens were diss ected to count retained eggs. The wings on the microscope slides were photographe d with a standard size referen ce (a length of steel measured to the thousandth of a millimeter with a caliper ). The photographs were opened in SigmaScan Pro 5 (Systat Software, Inc., Point Richmond, CA ), calibrated for size, and measured from the alular notch to th e distal end of R2, excluding fringe hairs (Packer and Corbet 1989). It was decided to use the distal end of R2 as a measurement point as it is an unambiguous standard landmark. Other studies have used the distal end of the wing, or some kind of other subjective measure of the maximal distance (Packer and Corbet 1989, Lima et al. 2003). While others have suggested transforming the winglengths thus ob tained by cubing the linear measure (Briegel

PAGE 70

70 1990 a), the recommendations of Siegel were followed here (Siegel et al. 1992), thus winglengths were not transformed. Analyses Regression analyses were conducted using S-Plus 7.0 for Windows (Insightful 2005). Analysis of Covariance was conducted using SAS 9.0 (SAS Institute). Fecundity was scored as the number of eggs in the egg raft plus the num ber of retained eggs in the killed females, provided these numbered fewer than 50. If female s retained more than 50 eggs, they were discarded from the analysis, as they became di fficult to count in high numbers. Only two females in this study retained more than 50 eggs. Simple linear regressions of each predictor on fecundity were performed to compare with published results of these pr edictors on fecundity for Culex quinquefasciatus females. A multiple linear regression was performe d on untransformed variables regressing fecundity on body size, bloodmeal size and age. Non-significant interactions and main effects were discarded in a stepwise fashion. Norm ality was checked with a Kolmogorov-Smirnov Test of Composite Normality, and homoscedasticity with a plot of residuals versus fitted values. A multiple regression of fecundity on the standa rdized values (Z-scores) of the predictors was also performed. Z-scores we re calculated by subtracting the mean value of a regressor from each observation, then dividing this by the standard deviation of the regressor. This produces mean values of zero and standard deviations of one. When used in multiple regression analysis, it allows a more standard interp retation of slope values, i.e. It apportions mean changes in response due to predictor variat ions of one standard devia tion (Marquardt 1980). Doing so allows one to order the predictors in terms of influence on a common scale.

PAGE 71

71 Significant interactions with age were explor ed with an ANCOVA testing for differences in slope of hematin or winglength on fecundity associated with the different age groups. A regression of arcsine square root transf ormed percentage hatch on age was performed, but every transformation of the data failed to prod uce normal residuals. Further analysis of this relationship was conducted with a Kruskal-Wallis test (Milton 1992). A regression of raw percentage hatch against age was performed to ge t a rough estimate of the effect of age on %age hatch. Results Winglength of female Culex quinquefasciatus ranged from 2.77 to 3.39 mm (Table 5-1). Amounts of hematin excreted and fec undity also varied. (Table 5-1). Significant linear regressions were found for all three factors, as well as the multiple regressions (Table 5-2, Figures 5-1, 5-2, 5-3). The fit of the full multiple model was considerably better than for any of the one-factor linear regressions. A ll individual parameters influenced fecundity with directions predicted by previous work, excep t for the unstandardized multiple model (Table 5-2). The standardized multiple model indicates the expected direction of influence of these parameters on fecundity (hematin positive, winglength positive, age negative). In the formation of the multiple models, it was discovered that there was a significant interaction between age and body size predicting fecundity. Analysis of this interaction by ANCOVA (slopes for winglength at varying levels of age) sh owed that in the younger age classes, slopes were positive and significant, (Table 53), but after 17 days of age, slopes were not significantly different from zero. For illust ration, significant slopes were grouped, and slopes not different from zero were grouped (Fig. 5-4).

PAGE 72

72 The Kruskall-Wallis test showed significant di fferences in %age hatch between the six age groups (Chi-square = 20.7657, df = 5, p-value = 0.0009) but an examination of the scatterplot of percentage hatch versus age failed to show a reliable tr end with age (Figure 5-5). Discussion As expected, both bloodmeal size and body size had a positive influenc e on fecundity. The degree of dependence on these two factors considered alone is rela tively low, but taken together with age, a very precise model explaining most of the variance in fecundity was achieved. Age was the factor that explained the greatest variance in fecundity, e ither as a single factor in the simple linear regressions, or as a factor in the standardized multiple regression model. The decline in fecundity with age had the greatest influence on fecundity in the standardized multiple model, indicating that a sta ndard deviation in age has a greater effect than a standard deviation in bloodmeal size or body si ze. That there is a noticeable decline in fecundity with age over a span of 25 days post-eclosion accords with one of the expectations of the antagonistic pleiotropy hypothesi s of aging, namely that age-a ssociated fitness effects are expected to be large in taxa experien cing high per-diem mortality (Williams 1957). Consideration of the two full models illustra tes the potentially confusing nature of unstandardized factors used as predictors in multiple regression models. The unstandardized form seems to qualitatively differ with respec t to the expected sign (+/-) for age. The explanation for this is that f actors were measured on different scales (Marquardt 1980). Each regression would perform identically when used as a predictive equation, but more intensive analysis is better accomplished by examining th e model with standardized regressors. In the standardized model, each predictor acco rds with the expected direction of influence on fecundity: i.e. Negative for age, positiv e for bloodmeal size and body size. Since the predictors are considered at equi valent intervals, one can judge th e relative contributions of each

PAGE 73

73 to the overall model. On the whole, age ha d the strongest influence on fecundity, followed by bloodmeal size and body size. The estimate for the decline in fecundity associ ated with age in the unstandardized simple linear regression of age on fecundity is similar to that reported for a wild Vero Beach strain of this species (Walter and Hacker 1974), but grea ter than those reported for Asian populations of this species (Walter and Hack er 1974, Suleman and Reisen 1979). The influence of age on fecundity may also be modified by temperature and humidity, so direct comparisons between these studies are difficult to make, as these other studies held the animals under different conditions.. It should be noted that the effect of age-rela ted declines in fecundity are only one measure of the impact of age on population growth in an age-structured population. Mosquitoes have high rates of daily mortality (Dow 1971, Reisen et al. 1991), and this is likely to account for a greater decline in fitness at the population level than cha nges in fecundity (Dye 1984). Nonetheless, the effect of these age-related f ecundity declines is substantial and should be incorporated into the construc tion of models of age-structur ed population growth wherever possible (Gotelli 1998). The interaction between adult body size and ag e accounted for a portion of the variance and increased the R2 of the model moderately. It also led to a more careful consideration of the effect of body size across the different age gr oups. The effect of body size on fecundity was not constant: in the younger groups, it was significant and positive wher eas in the older groups it had no significant influence. A possi ble explanation would be that if mobilization of teneral protein reserves is interrupted as a consequence of age, or if most of these reserves have disappeared, then the only significant factor predic ting fecundity would be bloodmeal size.

PAGE 74

74 Briegel demonstrated with Aedes aegypti the somewhat surprising fact that larger females invested progressively less lipid into their yolk mass with in creasing gonotrophic age, despite having access to sucrose from which to synthesize lipids. This was not seen in smaller females (Briegel et al. 2002). In fact, in both groups, total lipid increased w ith age. Associated with this finding was an observation that the slope of the f ecundity versus age curve was steeper for larger females than it was for smaller females, indicatin g that proportionate to their body size, larger females tended to produce less eggs in later ovipos itions than in earlier ones. If such a dynamic also exists in Culex quinquefasciatus then this could show up as an interaction between age and body size predicting fecundity. Another possible explanation for this might be a d ecline in protein re serves brought about by the sugar-only diet of the females prior to bloo dfeeding (Lang et al. 1965). If protein reserves are totally exhausted, then they cannot positively in fluence fecundity. If we also postulate that there is no or low correlation between body size and bloodmeal size, then any positive effect of body size on fecundity would disapp ear when teneral reserves are exhausted. In this dataset, only 12.9% of the variance in bl oodmeal size can be apportioned to differences in body size (data not shown). The relationship between teneral protein rese rves and fecundity is best seen with autogenous mosquitoes, as protein fo r the first egg batch is derived entirely from teneral reserves (Telang and Wells 2004). Because in this study the correlation between body size and bloodmeal size is low (larger females did not take significantly larger bloodmeals), there is no discernable advantage to being larger. This is the first time that an interac tion between adult mosquito body size and age predicting fecundity has been re ported, and it should stimulat e further research into the

PAGE 75

75 possibility of interactions with age in predic tive models of mosquito fecundity. Particular consideration should be given to generating a wi de variety of sizes of mosquito, as well as perhaps greater diversity of ag e groups. This may give a be tter idea of the timing of any interaction effects. Other experiments should al so be performed testing whether this type of interaction between body size and age is evid ent in females given multiple bloodmeals.

PAGE 76

76 Table 5-1. Summary statistic s for factors and responses used in regression models. Statistic Hematin (g) Winglength (mm) AgeFecundity (# of eggs in raft) Retained eggs percentage hatch Min 7.47 2.775430 1.15 Mean 16.87 3.0612.141511.27 81.47 Max 29.39 3.392523135 100 N 130 130130130130 130 Std. Deviation 4.04 0.116.9844.174.58 18.58

PAGE 77

77 Table 5-2. Summaries of five separate regression equatio ns predicting fecundity of Culex quinquefasciatus fecundity from individual factors and combinations of factors. Regressors Parameter Coefficient Std. Error DF T or F P R2 Intercept 222.616.513634.18 <0.01 Age -5.750.4901-11.72 <0.01 Model 30.241,128137.40 <0.01 Age 0.52 Intercept 72.8514.794.93 <0.01 Hematin 4.74 0.855.56 <0.01 Model 1,12830.95 <0.01 Hematin 0.19 Intercept -123.88102.71-1.21 0.23 Winglength 90.3133.492.70 <0.01 Model 1,1287.27 <0.01 Winglength 0.05 Intercept -409.13 137.48-2.98 <0.01 Hematin 3.600.546.71 <0.01 Winglength 187.0345.594.10 <0.01 Age 24.2010.792.24 <0.05 Winglength*Age -9.813.53-2.78 <0.01 Model 4,12585.04 <0.01 Full model, unstandardized 0.73 Intercept 153.502.0276.13 <0.01 Hematin 14.542.176.71 <0.01 Winglength 7.562.183.47 <0.01 Age -31.902.04-15.64 <0.01 Winglength*Age -5.942.14-2.78 <0.01 Full model, standardized Model 4,12585.04 <0.01 0.73 *The predictors and interactions that were not significant at the =0.05 level were removed as part of a stepwise model reduction. Indicates degrees of freedom for overall model F-test.

PAGE 78

78 510152025Age (Days post-eclosion) 0 50 100 150 200Fecundity (Eggs laid + retained) Fecundity= 222.61 5.74(Age) Figure 5-1. Scatterplot of f ecundity versus unstandardized age at bloodfeeding for Culex quinquefasciatus.

PAGE 79

79 51015202530Hematin Excreted ( g) 0 50 100 150 200Fecundity (eggs laid + retained) Fecundity=72.85 + 4.74 (Hematin) Figure 5-2. Scatterplot of fecundity versus unstandardized bloodmeal size for Culex quinquefasciatus

PAGE 80

80 2.72.93.13.3Winglength (mm) 0 50 100 150 200Fecundity (eggs laid + retained) Fecundity=-123.88 + 90.31(Winglength) Figure 5-3. Scatterplot of fecund ity versus unstandardized body si ze for Culex quinquefasciatus.

PAGE 81

81 2.72.93.13.3 0 50 100 150 200 250 F e c u n d it y ( e g g s la id + r e t a in e d )Win g len g th ( mm ) Figure 5-4. Scatterplot of uns tandardized winglength versus fecundity showing the response between ages 513 days (solid line, circle s) and 17-25 days (br oken line, triangles). Table 5-3. ANCOVA analysis desc ribing slope of the fecundity versus winglength relationship for 6 Culex quinquefasciatus age classes. Also shown are the slopes of the pooled young versus old mosquitoes. Note that in both cases, after 13 days, the slopes are not significantly different from zero. Age (Days) Parameter EstimateStd. Errorn T P 5 Winglength 151.04341.48827 3.64 <0.01 9 Winglength 115.00745.11632 2.55 0.0121 13 Winglength 175.78536.79632 4.78 <0.01 17 Winglength 10.51648.54524 0.22 0.8289 21 Winglength -109.853103.16311 -1.06 0.2891 25 Winglength 54.767129.1884 0.42 0.6724 Ages 5-13 Winglength 131.77327.74391 4.75 <0.01 Ages 17-25 Winglength -6.31749.09039 -0.13 0.8978

PAGE 82

82 Figure 5-5. Scatterplot of Per centage hatch versus age for Culex quinquefasciatus

PAGE 83

83 CHAPTER 6 DISCUSSION Fecundity is an important parameter in the overall fitness of female organisms, as well as males, albeit indirectly. At the popul ation level, the mean number of offspring produced by females has an effect on the growth characteristics of the population. Anautogenous mosquitoes, like many inse cts, are capable of producing a great number of eggs in their lifetime. This reproductive output is pr edicated upon several factors, such as larval nutrition (Akoh et al. 1992), adult nutrient acquisition (bloodfeeding) (Akoh et al. 1992, Briegel 2003), and survival. Another factor which is known to affect reproductive rate in many organisms is age (Akoh et al. 1992, Mahmood et al. 2004). The gradual breakdown in physiological competence, and associated decreases in performance as an organism ages is referred to as senescence. The studies described in this work presen ted here examined the role of bloodmeal size, body size and age in dete rmining the fecundity of Culex nigripalpus and Culex quinquefasciatus Body size served as the proxy for la rval nutrient ac quisition, since the two factors have been found to be correla ted (Akoh et al. 1992, Koella and Offenberg 1999). Larval Nutrition The effects of mosquito larval nutrition on reproduction have been well studied in autogenous species. In autogenous species, the entire protein input into the yolk of the eggs is of larval origin (Clements 2000 a) a nd no bloodfeeding is requ ired, at least for the first gonotrophic cycle. Lipid and carbohydrat e may be larvally-derived, or may be acquired from sugar feeding.

PAGE 84

84 Anautogenous species (species requiring bl ood for vitellogenesis) likewise realize benefits from the larval stage, both from th e teneral reserves acquired (Briegel 2003), and also because greater larval nutrition correlate s with greater adult body size, which in turn is correlated with a larger number of ovarioles in the ovari es (Clements 2000 c). Larger mosquitoes are also able to take larg er bloodmeals (though the strength of this relationship varies). This means that larger mosquitoes are capable of greater protein acquisition. The effects of larv al nutrition are seen as a si mple correlation between body size and fecundity in the species in ques tion (Briegel 1990 a, Akoh et al. 1992, Lyimo and Takken 1993, Armbruster and Hutc hinson 2002, Lima et al. 2003). In the series of experiment s reported here, there was consistently a significant effect of body size on fecundit y, indicating that larval nut rient reserves and adult body size are important predictors of fecundity in Culex nigripalpus and Culex quinquefasciatus This result agrees with results obtained for many mosquito species (Nasci 1986, Briegel 1990 b, Akoh et al 1992, Lyimo and Koella 1992, Lyimo and Takken 1993, Xue et al. 1995, Armbruster a nd Hutchinson 2002, Lima et al. 2003). Bloodmeal Size The size and quality of the bloodmeal is a nother important determinant of fecundity in anautogenous mosquitoes. The bloodmeal provides the only protein input for egg production aside from protein acquired in th e larval stage. Blood also furnishes substantial amounts of lipid and carbohydrate, which are used in vitellogenesis. In most studies to date, the size of the bloodmeal has been shown to be positively correlated with the number of eggs produced (Cochrane 1972, Edman and Lynn 1975, Clements 2000 a, Briegel 2003). This has been shown to be the case in both Culex nigripalpus (Edman and Lynn 1975), and Culex quinquefasciatus (Akoh et al. 1992).

PAGE 85

85 The results of the experiments reported he rein confirm these results. They also provide a precise quantification of the numbe r of eggs produced for a given amount of voided hematin. The strength of these relationshi ps varied by species, and also with age. This provides a baseline estimate of the st rength of the dependence of fecundity on bloodmeal size in these two species of mo squitoes, as well as how bloodmeal size combines with other factors to influence f ecundity. This knowledge can be used to better understand the role of life history events (such as bloodfeeding a nd larval feeding) shaping the habits of these species. Age The process by which the health and vitality of an organism diminish with age is known as senescence, and it is found to varying degrees in ma ny taxa. What is less well understood are the processes which contribute to this decline in facultie s as an organism ages. There are some evolutionary theories re garding the origin and maintenance of senescence in natural populati ons (Gavrilov and Gavrilova, 2002). The major predictions of these theories are that s hort lived organisms are expected to experience the negative physiological consequences of aging more severe ly and at an earlier age than longer lived organisms, due to the weakness of selec tion against late-life fitness declines. The phenomenon of senescence has implicati ons not only for the aging individual, but also collectively for the population as a whole, when cohorts of organisms become old-biased. Senescence is a plastic trait unde r natural selection, and can be influenced by artificial selection in relativ ely few generations (Stearns et al. 2000 a). This helps to confirm an evolutionary e xplanation for senescence.

PAGE 86

86 Mosquitoes are known to have high extrin sic mortality (Dow 1971, Reisen et al. 1991, Elizondo-Quiroga et al. 2006), and as such have likely been selected for high earlylife fecundity (an extrapolation from the theo ries mentioned above). This means that reproductive capacity should reach its maximum shortly after the animal reaches sexual maturity (4 days post-eclosion for many mosqu ito species) and decline thereafter. This has been demonstrated in several mosquito species (Jalil 1974, Walter and Hacker 1974, Gomez et al. 1977, Suleman and Reisen 1979, Akoh et al. 1992, Mahmood et al. 2004), but the rate at which fecundity declines can be specific to the species or the geographic strain of the species in ques tion (Walter and Hacker 1974). In disease vectors such as mosquitoes, th e decline in fecundity with age may have epidemiologic significance if it serves to d ecrease the abundance of vectors available for enzootic or epidemic disease transmission over the course of a transmission season. Arbovirus transmission is in part dependent in part, on vector abundance, and if abundance is not sufficient, transmission is unlikely (Mitchell et al. 1980, 1983). It should be noted, however, that small populati ons of aged mosquitoes can sometimes sustain epidemic transmission (Day 2001). The experiments reported herein provide es timates as to the effect of age on the fecundity of Culex nigripalpus and Culex quinquefasciatus from Florida. The demonstration of an age-re lated fecundity decline in Culex nigripalpus is the first such reported for this species. The aging experiments in this work were conducted in a manner that did not allow repeated bloodfeeding, rather the bloodmeals gi ven represented the first that any female had in her lifetime. It is not clear from th ese results whether the effects of senescence

PAGE 87

87 might have been ameliorated had the females had access to blood on a regular basis. Nonetheless, the slopes of the age related fecundity declines are similar to those of other Culex species studied (W alter and Hacker 1974, Mahmood et al. 2004). Interactions An interesting and novel result to emerge fr om this work is the demonstration of interactions between bloodmeal size (Chapter 4), and body size (Chapter 5) with age. This is the first time such a phenomenon has been investigated, and bears closer consideration. In the case of the interaction between bloodmeal size and age predicting fecundity in Culex nigripalpus the interaction was in large part due to the re sponse of fecundity to bloodmeal size in the last two age groups ( 20 and 25 days old). The sample size for estimating this effect was low, but the slope estimate was large and negative. A possible explanation for the result is th at digestive or vitellogenic pr ocesses are impaired in older females. In the case of the interact ion between body size and age in Culex quinquefasciatus the result was more robust due to a larger sample size. A simple physiological explanation suggests itself: that teneral reserves of protein are depleted as the mosquito ages and thus at older ages, increased body si ze fails to result in increased fecundity. Another explanation might be offered by the finding of Briegel, working with Aedes aegypti : that larger females invest progressiv ely less lipid into cogenesis as they age, while smaller females do not (Briegel et al. 2002). Differentia ting between these two hypotheses could only be accomplished by car eful biochemical analyses of the mosquitoes as they age.

PAGE 88

88 It is unknown if this interaction effect w ould disappear if repeated bloodmeals had been made available. If bloodfeeding had b een allowed at regular intervals, perhaps metabolically-depleted teneral protein re serves would be replenished, although in general, bloodmeal protein is generally onl y used for egg production (Briegel 2003). Future Research The results of these experiments suggest some new avenues and techniques for research into factors govern ing fecundity in mosquitoes. The interaction terms discovered in Chapte rs 4 and 5 suggest that there may be some important physiological cha nges in blood digestion or levels of teneral reserves that occur in Culex mosquitoes with age. With refere nce to the interaction between body size and age predicting fecundity in Culex quinquefasciatus the role of tene ral protein in egg production should be examined. It is known th at the entire protein input necessary for first-cycle egg production in au togenous mosquitoes, including Culex pipiens molestus Forskal, are derived from larval nutriti on (Clements 2000 c). It is not therefore unreasonable to suspect that teneral prot ein also contributes to reproduction in anautogenous species such as Culex quinquefasciatus If so, the deple tion of such protein reserves in aging adult mosquitoes could be examined. Different taxa and populations could then be compared in order to determ ine the relative contributions of teneral and bloodmeal-derived protein inputs used for re production. This would be of interest because it would allow a partitioning of th e effects of larval and adult nutrition on lifetime fitness. With reference to the interaction be tween bloodmeal size and age predicting fecundity detected in Culex nigripalpus (Chapter 4), the func tioning of digestive and oogenic processes should be examined in aging mosquitoes. If these become

PAGE 89

89 compromised in some way, then the interact ion could be explained, and a mechanism of senescence in this species could be elucidated. Aging and senescence are not biologically fi xed traits, but rather have been shown to be modified by selection (Stearns et al. 2000 b). It is entirely reasonable to suspect that different species and diffe rent populations within a speci es have been selected for different lifespans. This is perhaps one reas on why the decline in fecundity with age has been shown to differ markedly between populat ions of mosquitoes (Walter and Hacker 1974). For this reason, life history traits such as senescence need to be examined with reference to a particular population. For exam ple, if one were to incorporate an agerelated fecundity reduction into an age-stru ctured population growth model, one would need to be careful to have parameter estimates derived from the population under consideration. If not, then the model and the field reality may differ greatly. One methodological consideration that these results suggest is that multiple factors should be included together in models pred icting fecundity as well as experiments to determine factors affecting fecundity. This allows for greater accura cy in prediction, and also allows some comparison of the re lative importance of the factors under consideration. It should also be obvious that the standardiz ation of parameters improves the interpretability of regr ession models, especially when higher order terms or interactions are present. The dangers of working with unstandardized predictors in regression models have been stated before (Marquardt 1980), name ly that significant effects could be unjustifiably discarded. From a theoretical standpoint, it is obvi ous that possible interactions should be sought out in multiple models, and if present an alyzed further. An interaction indicates

PAGE 90

90 that the response to one factor changes at diffe rent levels of the inte racting factor. This may be seen as a complication, but in realit y, inclusion of signifi cant interaction terms can often better describe the relationship between multiple predictors and a response. The results of these experiments shed some new light onto the interplay of factors affecting fecundity in two Culex species. Continued investigation into the factors governing reproductive success of mosquitoes will provide interesting insights into the natural history of these fascin ating and important insects.

PAGE 91

91 APPENDIX A WING MEASUREMENT VALIDATION Objective To test and validate a method using photographs and image anal ysis software for precisely measuring mosquito winglength. Methods Ten pieces of aluminum from a soft drink can were cut to random sizes and measured to the nearest thousandth of a millimeter with a caliper. The pieces were numbered and measurements were recorded. The metal piece s were then photographed adjacent to a 6.758mm reference. The photographs were imported into SigmaScan (Systat Software Inc. Richmond CA) and calibrated using the SigmaScan 2-point recalibration functi on. The metal pieces were then measured using the SigmaScan default 2-point measurement system in the program and the measurements for each were stored. A linear regression predicting callipered m easurement from the SigmaScan measurement was performed using S-Plus (Insightful Corp. Seattle WA). Results The linear regression was highly si gnificant (Fig. A-1) and had an intercept near zero and a slope that approached one (Table A-1). The inte rcept was not significant, and thus cannot be distinguished from zero. Conclusion This method of measurement is accurate, precise and easy to use. It can easily be used to measure between two distinct landmarks on a mo squito wing mounted on a microscope slide (Fig. 2-1).

PAGE 92

92 02468101214SigmaScan Measurement (mm) 0 4 8 12Caliper Measurement (mm) Figure A-1. Regression plot show ing relationship between caliper measurement and SigmaScan measurement.

PAGE 93

93 Table A-1. Results of linear regression analysis of callipered length against length measured in SigmaScan. Parameter Estimate Standard ErrorP Intercept 0.05420.06930.4573 Slope 1.00210.0101<0.0001

PAGE 94

94 APPENDIX B HEMATIN STANDARD CURVE Introduction Hematophagous insects void several products in their feces following a blood meal. These products include uric acid deri ved from tubal fluids (Briegel 1980, Clements 2000 d), as well as products of blood digestion follo wing a bloodmeal (Clements 2000 d). One of these products is hematin, a degraded digestion product of hemogl obin. It was suggested by Briegel (1980) that quantifying the hematin content of excrement might allow one to estimate the relative size of a mosquito bloodmeal. The method of quantifica tion involves reading th e absorbance of the sample at a wavelength of 387 nm, the absorpti on peak for hematin. Figure B-1 illustrates a sample spectral scan of Culex quinquefasciatus excrement showing the absorption peak corresponding to the maximum for hematin. Many authors have used this approach successfully to qua ntify bloodmeals their experimental animals consume (Briegel 1980, Mitchell and Briegel 1989, Briegel 1990 a, Hogg et al. 1996, Ferguson et al. 2003). The chief advantage of this method of bloodmeal quantification is that it does not involve intrusive and laborious methods of anesthetic use on experimental animals and weighing of animals before and after bloodfeeding. Objective To develop a regression model that accurately calculates the amount of hematin in mosquito excrement, as a means of quan tifying the relative size of the bloodmeal. Methods Three stock solutions of 1.00 mg/ml porcine he matin (MP Biomedicals, Irvine CA) in 1% lithium carbonate were prepared using an analytical balance. From these stock solutions, 14

PAGE 95

95 dilutions were prepared, from 0-26g/ml at in tervals of 2.00 g/ml. The absorbance of these solutions at 387nm (the peak absorbance for he matin) was measured and the data recorded. Absorbance at 387nm was regressed on hematin c oncentration using a linear regression in S-Plus 7.0 for Windows (Insightful Corp.). The slope estimate for this regression was used in all calculations of hematin conc entration used in this Thesis. Results A regression equation was constructed to predict hematin mass from absorbance. Conclusion This technique allows one to quantify the hema tin present in a sample of excreta, thereby providing an indirect estimate of mosquito bloodmeal sizes.

PAGE 96

96 Figure B-1. Example of a spectral scan of Culex quinquefasciatus excrement. Note peak corresponding to maximal absorbance for hematin at 387nm.

PAGE 97

97 0.30.81.31.8 387 0 5 10 15 20 25Hematin Concentration ( g/ml) Figure B-2. Standard curve used for calculating amount of hematin in mosquito excreta. R2 for this regression was 0.9969, F=12830 on 1 and 40 degrees of freedom. N=42 Table B-1. Results of the regre ssion analysis for absorbance ve rsus concentration. The slope estimate was used to calculate all hematin quantities in this work. Parameter EstimateStandard ErrorP Intercept 0.05560.13460.6817 Slope 14.3660.1268<0.01

PAGE 98

98 APPENDIX C MOSQUITO WINGLENGTH VS. WEIGHT REGRESSIONS Introduction Several studies have show that a repeatab le, meaningful measure of overall size of a mosquito is the length of the wing (Briegel 1990 a, Nasci 1990, Siegel et al. 1992, Hogg et al. 1996, Koella and Offenberg 1999, Armbruster and Hutchinson 2002, Lima et al. 2003). This is usually accomplished by measuring the wing from the alular notch to the wingtip, excluding fringe hairs. Objective To determine the relationships between teneral weights of Culex quinquefasciatus and Cx. nigripalpus females and their winglengths. The purpos e of this was to find a proxy measurement for female size for use in regression models predicting the fecundity of these species. Materials and Methods Larvae were reared in enameled metal pans measuring 24 x 36 x 5cm containing approximately 700 ml of tap water. Pans we re set with 3 egg rafts each of colonized Culex quinquefasciatus or Culex nigripalpus Food was provided daily to each pan as 20 ml of slurry containing 20 mg/ml 1:1 Brew ers yeast/liver powder. Forty-two female pupae of Culex quinquefasciatus and 43 female pupae of Culex nigripalpus were removed individually, a nd dabbed dry with a paper towel. Pupae were weighed on a Cahn Millibalance Model 7500 to the nearest 100th of a milligram. Following weighing, pupae were placed in individual 40 mm vials with 4 ml tapwater and allowed to emerge. Following emergence, the mosquitoes were kille d and their wings removed and adhered to a microscope slide with clear double-sided tape. A coverslip was placed over the wings to prevent damage. Wings were photographed ne xt to standard size reference and the distance between the

PAGE 99

99 alular notch and the end of wing vein R2 (excluding fringe setae) wa s measured using SigmaScan (Appendix 1). Data were entered into S-Plus and linear regressions performed predicting mass from winglength. Significant linear relationshi ps were found between pupal mass and winglength in both species (Table B-1, Figs. B-1, B-2) Slopes and intercepts for both regressions were quite similar, with that for Culex quinquefasciatus being slightly higher. Table C-1. Results of two regression analyses predicting female mass from winglength. All regressions were significant at the =0.05 level. Species Parameter Estimate Standard Error dfTPR2 Intercept -4.171.77-2.360.02 Cx. quinquefasciatus Winglength 2.53.56 1,40 4.54<0.01 0.34 Intercept -4.12.98-4.21<0.01 Cx. nigripalpus Winglength 2.33.32 1,41 7.37<0.01 0.57 Conclusion Winglength is an adequate predictor of pupal mass in Culex quinquefasciatus and Culex nigripalpus.

PAGE 100

100 2.93.03.13.23.3Winglength (mm) 3.0 3.5 4.0 4.5Mass (mg) Figure C-1. Linear regre ssion of winglength on wet mass of pupae of female Culex quinquefasciatus n=42

PAGE 101

101 2.82.93.03.13.23.33.4Winglength (mm) 2.1 2.6 3.1 3.6Mass (mg) Figure C-2. Linear regre ssion of winglength on wet mass of pupae of female Culex nigripalpus N=43

PAGE 102

102 LIST OF REFERENCES Agnew, P., C. Haussy, and Y. Michalakis. 2000. Effects of density and larval competition on selected life history traits of Culex pipiens quinquefasciatus (Diptera: Culicidae). J. Med. Entomol. 37: 732-735. Ahid, S. M., P. S. Vasconcelos, and R. Lourenco-de-Oliviera. 2000. Vector competence of Culex quinquefasciatus say from different regions of Brazil to Dirofilaria immitis Mem. Inst. Oswaldo Cruz. 95: 769-75. Akoh, J. I., F. I. Aigbodion, and D. Kumbak. 1992. Studies on the effect of larval diet, adult body weight, size of blood-meal, and age on the fecundity of Culex quinquefasciatus (Diptera:Culicidae). Inse ct Sci. Appl. 13: 177-181. Allan, S. A., U. R. Bernier, and D. L. Kline. 2006. Laboratory evaluation of avian odors for mosquito (Diptera: Culicidae) at traction. J. Med. Entomol. 43: 225-231. Armbruster, P., and R. A. Hutchinson. 2002. Pupal mass and wing length as indicators of fecundity in Aedes albopictus and Aedes geniculatus (Diptera: Culicidae). J. Med. Entomol. 39: 699-704. Blackmore, M. S., and C. C. Lord. 1994. Size-fecundity relationship in Aedes albopictus The Vector Cont. Bull. N.C. States 3: 90. Blackmore, M. S., and C. C. Lord. 2000. The relationship between size and fecundity in Aedes albopictus J. Vector Ecol. 25: 212-217. Boike, A. H. 1963. Observations on Culex nigripalpus Theobald in a typical hammock area of North Central Florida. Mosq. News. 23: 345-348. Bonduriansky, R., and R. J. Brooks. 1999. Reproductive allocati on and reproductive ecology of seven species of Dipter a. Ecol. Entomol. 24: 389-395. Bradshaw, W. E., and C. M. Holzapfel. 1992. Reproductive consequences of densitydependent size variation in the pitcherplant mosquito, Wyeomyia smithii (Diptera, Culicidae). Ann. Entomol. Soc. Am. 85: 274-281. Briegel, H. 1980. Determination of uric acid and hematin in a single sample of excreta from blood fed insects. E xperientia 36: 1428. Briegel, H. 1990 a. Fecundity, metabolism, and body size in Anopheles (Diptera: Culicidae), vectors of malaria. J. Med. Entomol. 27: 839-50. Briegel, H. 1990 b. Metabolic relationship between female body size, reserves, and fecundity of Aedes aegypti J. Insect Physiol. 36: 165-172. Briegel, H. 2003. Physiological bases of mosquito ecology. J. Vector Ecol. 28: 1-11.

PAGE 103

103 Briegel, H., and S. E. Timmermann. 2001. Aedes albopictus (Diptera: Culicidae): physiological aspects of development and reproduction. J. Med. Entomol. 38: 566-71. Briegel, H., M. Heft i, and E. DiMarco. 2002. Lipid metabolism during sequential gonotrophic cycles in large and small female Aedes aegypti J. Insect Physiol. 48: 547-554. Calisher, C. H. and N. Karabatsos. 1988. Arbovirus serogroups: definition and geographic distribution, pp. 19-57. In T. P. Monath [ed.], The Arboviruses: Epidemiology and Ecology. CRC Press, Boca Raton. Chadee, D. D. and J.S. Haeger. 1986. A description of the egg of Culex (Culex) nigripalpus Theobald from Florida, with not es on five egg rafts (Diptera : Culicidae). Mosq. Syst. 18: 288-292. Charlesworth, B. 2000. Fisher, Medawar, Hamilton and the evolution of aging. Genetics 156: 927-931. Clements, A. N. 2000 a. Vitellognesis, pp. 360-379. In A. N. Clements [ed.], The Biology of Mosquitoes. CABI Publishing, New York. Clements, A. N. 2000 b. Adult Digestion, pp. 272-303. In A. N. Clements [ed.], The Biology of Mosquitoes. CABI Publishing, New York. Clements, A. N. 2000 c. Nutrition and fertility of an autogenous mosquitoes, pp. 408-423. In A. N. Clements [ed.], The Biology of Mosquitoes. CABI Publishing, New York. Clements, A. N. 2000 d. Adult diuresis, excretion and defaecation, pp. 304-326, The Biology of Mosquitoes. CABI Publishing, New York. Cochrane, A. 1972. Body weight and blood meal weight as factors aff ecting egg production of the tree-hole mosquito, Aedes triseriatus (Say). Proc. N.J. Mosq. Ext. Assoc: 65-78. Colless, D. H., and W.T. Chellapah. 1960. Effects of body weight a nd size of blood-meal upon egg production in Aedes aegypti (Linnaeus) (Diptera, Culicidae). Ann. Trop. Med. Parasitol. 54: 475-482. Darlington, R. B. 1990. Path analysis and hierarchical designs, pp. 170-190, Regression and Linear Models. McGraw-Hill, New York. Darsie, R. F. and R.A. Ward. 2004. Identification and Geogra phical Distribution of the Mosquitoes of North America, North of Mexic o. University of Florida Press, Gainesville. Day, J. F. 2001. Predicting St. Louis encephalitis virus ep idemics: lessons from recent, and not so recent, outbreaks. Annu. Rev. Entomol. 46: 111-38.

PAGE 104

104 Day, J. F. 2005. Host-seeking strategies of mosquito disease vectors. J. Am. Mosq. Control Assoc. 21: 17-22. Day, J. F., G. A. Curtis, and J. D. Edman. 1990 a. Rainfall-directed oviposition behavior of Culex nigripalpus (Diptera: Culicidae) and its infl uence on St. Louis encephalitis virus transmission in Indian River County, Florida. J. Med. Entomol. 27: 43-50. Day, J. F., A.M. Ramsey, and J.T. Zhang. 1990 b. Environmentally mediated seasonal variation in mosquito body si ze. Environ. Entomol. 19: 469-472. Day, J. F. and G.A. Curtis. 1989. Influence of rainfall on Culex nigripalpus (Diptera:Culicidae) blood-feeding behavior in I ndian River County, Florida. Ann. Entomol. Soc. Am. 82. Dow, R. P. 1971. The dispersal of Culex nigripalpus marked with high concentrations of radiophosphorus. J. Med. Entomol. 8: 353-363. Dow, R. P., K.E. Coleman, and T.H Work. 1964. Isolation of St. Loui s encephalitis viruses from mosquitoes in the Tampa Bay area of Florida during the epid emic of 1962. Am. J. Trop. Med. Hyg. 13: 462-468. Dye, C. 1984. Models for the population dynamics of the Yellow Fever Mosquito, Aedes aegypti J. Anim. Ecol. 53: 247. Edman, J. D., and H. C. Lynn. 1975. Relationship between blood meal volume and ovarian development in Culex nigripalpus (Diptera: Culicidae). En tomol. Exp. Appl. 18: 492496. Elizondo-Quiroga, A., A. Flores-Suarez, D. Elizondo-Quiroga, G. Ponce-Garcia, B. J. Blitvich, J. F. Contreras-Cordero, J. I. Go nzalez-Rojas, R. Mercado-Hernandez, B. J. Beaty, and I. Fernandez-Salas. 2006. Gonotrophic cycle and survivorship of Culex quinquefasciatus (Diptera: Culicidae) using sticky ovitraps in Monterrey, northeastern Mexico. J. Am. Mosq. Control Assoc. 22: 10-14. Ferguson, H. M., M. J. Mackinnon, B. H. Chan, and A. F. Read. 2003. Mosquito mortality and the evolution of malaria vi rulence. Evolution. 57: 2792-804. Fernandes, L., and H. Briegel. 2005. Reproductive physiology of Anopheles gambiae and Anopheles atroparvus J. Vector Ecol. 30: 11-26. Fitt, G. P. 1990. Comparative fecundity, clutch si ze, ovariole number and egg size of Dacus tryoni and D. jarvisi and their relationship to body size. Entomol. Exp. Appl. 55: 11. Gavrilov, L. A., and N. S. Gavrilova. 2002. Evolutionary theories of aging and longevity. Sci. World J. 2: 339-356.

PAGE 105

105 Gomez, C., J. E. Rabinovich, and C. E. Machado-Allison. 1977. Population analysis of Culex pipiens fatigans Wied. (Diptera: Culicidae) unde r laboratory conditions. J. Med. Entomol. 13: 453-63. Gotelli, N. J. 1998. Age-Structured Population Growth, pp. 49-80. In N. J. Gotelli [ed.], A Primer of Ecology, 2 ed. Sinaue r Associates, Sunderland, MA. Hair, J. A. 1968. Observations on two species of Culex of the subgenus Melanoconion Mosq. News. 28: 425-429. Hall, M. H., S. M. Dutro, and M. J. Klowden. 1990. Determination by near-infrared reflectance spectroscopy of mosquito (Dipte ra: Culicidae) bloodm eal size. J. Med. Entomol. 27: 76-9. Hogg, J. C., M. C. Thomson, and H. Hurd. 1996. Comparative fecundity and associated factors for two sibling species of the Anopheles gambiae complex occurring sympatrically in The Gambia. Med. Vet. Entomol. 10: 385-91. Huitema, B. A. 1980. Analysis of heterogeneous regres sion case: Johnson-Neyman technique, pp. 270-297. In B. A. Huitema [ed.], The Analysis of Covariance and Alternatives. John Wiley & Sons, New York. Hurd, H., J. C. Hogg, and M. Renshaw. 1995. Interactions between bloodfeeding, fecundity and infection in mosquitoes Parasitol. Today 11: 411-416. Insightful. 2005. S-Plus Guide to Statistic s. Insightful Corporation. Jalil, M. 1974. Observations on the fecundity of Aedes triseriatus (Diptera: Culicidae). Entomol. Exp. Appl. 17: 223. Kirkwood, T. B., and M. R. Rose. 1991. Evolution of senescence: la te survival sacrificed for reproduction. Philos. Trans. R. Soc. Lond. B Biol. Sci.. 332: 15-24. Knight, K. L., and Pugh, R.B. 1973. A bibliography of the mosquito publications of Fred. V. Theobald (1868-1930). Mosq. Syst. 5: 230-234. Koella, J.C. and J. Offenberg. 1999. Food availability and parasite infection influence the correlated responses of life hi story traits to selection for age at pupation in the mosquito Aedes aegypti J Evol. Biol. 12: 760. Lang, C. A., H. Y. Lau, and D. J. Jefferson. 1965. Protein and nucleic acid changes during growth and aging in the mos quito. Biochem. J. 95: 372-377. Levenrook, L. 1986. Protein synthesis in relation to insect aging: An overview, pp. 201-216. In Collatz, K.G. and R.S. Sohal [eds.], In sect Aging. Springe r-Verlag, New York.

PAGE 106

106 Lima, C. A., W. R. Almeida, H. Hurd, and C. M. Albuquerque. 2003. Reproductive aspects of the mosquito Culex quinquefasciatus (Diptera:Culicidae) infected with Wuchereria bancrofti (Spirurida: Onchocercidae). Mem. Inst. Oswaldo Cruz 98: 217-22. Lounibos, L. P., V. L. Larson, and C. D. Morris. 1990. Parity, fecundity and body size of Mansonia dyari in Florida. J. Am. Mosq. Control Assoc. 6: 121. Lowrie, R. C., Jr., M. L. Eberhard, P. J. Lammie, C. P. Raccurt, S. P. Katz, and Y. T. Duverseau. 1989. Uptake and development of Wuchereria bancrofti in Culex quinquefasciatus that fed on Haitian carriers with different microfilaria densities. Am. J. Trop. Med. Hyg. 41: 429-35. Lyimo, E. O., and J. C. Koella. 1992. Relationship between body size of adult Anopheles gambiae s.l. and infection with the malaria parasite Plasmodium falciparum Parasitology 104: 233-7. Lyimo, E. O., and W. Takken. 1993. Effects of adult body size on f ecundity and the pre-gravid rate of Anopheles gambiae females in Tanzania. Med. Vet. Entomol. 7: 328-332. Mahmood, F., W. K. Reisen, R. E. Chiles, and Y. Fang. 2004. Western equine encephalomyelitis virus infection affect s the life table characteristics of Culex tarsalis (Diptera: Culicidae). J. Med. Entomol. 41: 982-986. Marquardt, D. W. 1980. A critique of some ridge regre ssion methods: Comment. J. Am. Stat. Assoc. 75: 87-91. Mather, T. N., and G. R. DeFoliart. 1983. Effect of host blood source on the gonotrophic cycle of Aedes triseriatus Am. J. Trop. Med. Hyg. 32: 189-193. McCall, P. J., and G. Eaton. 2001. Olfactory memory in the mosquito Culex quinquefasciatus Med. Vet. Entomol. 15: 197-203. Meyer, R. P. 1989. Estimation of vectorial capacity: pa thogen extrinsic incubation and vector competence. Bull. Soc. Vector Ecol. 14: 60-66. Meyer, R. P., J. L. Hardy, and S. B. Presser. 1983. Comparative vector competence of Culex tarsalis and Culex quinquefasciatus from the Coachella, Imperial, and San Joaquin Valleys of California for St. Louis encepha litis virus. Am. J. Trop. Med. Hyg. 32: 305-11. Milton, S. J. 1992. Kruskall-Wallis k-Sample test for location: unmatched data, pp. 437-439, Statistical Methods in the Biological and Health Sciences. McGraw-Hill, New York. Mitchell, C. J. 1983. Mosquito Vector Competence and Arboviruses, pp. 63. In K. F. Harris [ed.], Current Topics in Vector Rese arch. Praeger Publishers, New York.

PAGE 107

107 Mitchell, C. J., D.B. Francy, and T.P. Monath 1980. Arthropod Vectors, pp. 313-379. In T. P. Monath [ed.], St. Louis Encephalitis. Ameri can Public Health A ssociation, Washington DC. Miura, T. and R.M. Takahashi. 1972. The fecundity of Aedes nigromaculis in the laboratoryeffects of body weight and size of blood meal. Mosq. News32: 417-421. Nasci, R. 1986. Relationship between adult mosquito (D iptera: Culicidae) body size and parity in field populations. Environ. Entomol. 15: 874-876. Nayar, J. K. 1982. Bionomics and Physiology of Culex nigripalpus (Diptera: Culicidae) of Florida: An Important Vector of Diseas es. University of Florida, Gainesville. Nayar, J. K., and J. Sauerman. 1975. The effects of nutrition on survival and fecundity in Florida mosquitoes. Part 3. U tilization of blood and sugar for fecundity. J. Med. Entomol. 12: 220. Nayar, J. K., and D. M. Sauerman, Jr. 1977. The effects of nutrition on survival and fecundity in Florida mosquitoes. Part 4. Effects of blood source on oocyte development. J. Med. Entomol. 14: 167. Nayar, J. K., and J. W. Knight. 1981. Occurence of ovariolar di latations in nulliparous mosquitoes: Culex nigripalpus Mosq. News. 41: 281-288. Nayar, J. K., J. W. Knight, and S. R. Telford, Jr. 1998. Vector ability of mosquitoes for isolates of Plasmodium elongatum from raptors in Florida. J. Parasitol. 84: 542-546. Olagbemiro, T. O., M. A. Birkett, A. J. Mordue Luntz, and J. A. Pickett. 2004. Laboratory and field responses of the mosquito, Culex quinquefasciatus to plant-derived Culex spp. oviposition pheromone and the oviposition cue skatole. J. Chem. Ecol. 30: 965-76. Packer, M. J., and P. S. Corbet. 1989. Size variation and reproductive success of female Aedes punctor (Diptera: Culicidae). Ecol. Entomol. 14: 297-309. Provost, M. W. 1969. The natural history of Culex nigripalpus pp. 46-62, St. Louis Encephalitis in Florida. Florida State Board of Health Monograph Series Number. 12. Ramasamy, M. S., R. Ramasamy, B. H. Kay, and C. Kidson. 1988. Anti-mosquito antibodies decrease the repr oductive capacity of Aedes aegypti Med. Vet. Entomol. 2: 87-93. Reisen, W. K., M. M. Milby, R. P. Meyer, A. R. Pfuntner, J. Spoehel, J. E. Hazelrigg, and J. P. Webb, Jr. 1991. Mark-release-recapture studies with Culex mosquitoes (Diptera: Culicidae) in Southern California. J. Med. Entomol. 28: 357-371.

PAGE 108

108 Reiskind, M. H., E. T. Walton, and M. L. Wilson. 2004. Nutrient-dependent reduced growth and survival of larval Culex restuans (Diptera: Culicidae): laboratory and field experiments in Michigan. J. Med. Entomol. 41: 650-656. Renshaw, M., M. W. Service, and M. H. Birley. 1994. Size variation and reproductive success in the mosquito Aedes cantans Med. Vet. Entomol. 8: 179-186. Roitberg, B. D., and I. Gordon. 2005. Does the Anopheles blood meal-fecundity curve, curve? J. Vector Ecol. 31: 83-86. Rosen, L. 1986. The natural history of Japanese ence phalitis virus. Ann. Rev. Microbiol. 40: 395-414. Rutledge, C. R., J. F. Day, C. C. Lord, L. M. Stark, and W. J. Tabachnick. 2003. West Nile Virus infection rates in Culex nigripalpus (Diptera: Culicidae) do not reflect transmission rates in Florida. J. Med. Entomol. 40: 253-258. Sardelis, M. R., M. J. Turell, D. J. Dohm, and M. L. O'Guinn. 2001. Vector competence of selected North American Culex and Coquillettidia mosquitoes for West Nile virus. Emerg. Infect. Dis. 7: 1018-22. Shaman, J., J. F. Day, and M. Stieglitz. 2003. St. Louis encephalitis virus in wild birds during the 1990 south Florida epidemic: the importa nce of drought, wetting conditions, and the emergence of Culex nigripalpus (Diptera: Culicidae) to arboviral amplification and transmission. J. Med. Entomol. 40: 547-54. Siegel, J. P., R. J. Novak, R. L. Lampman, and B. A. Steinly. 1992. Statistical appraisal of the weight-wing length relationship of mo squitoes. J. Med. Entomol. 29: 711-4. Stearns, S. C., M. Ackermann, M. Doebeli, and M. Kaiser. 2000. Experimental evolution of aging, growth, and reproduction in fruitflies. Proc. Nat. Acad. Sci. USA: 97: 3309-3313. Suleman, M. R., and W.K. Reisen. 1979. Culex quinquefasciatus Say: life table characteristics of adults reared from wild-caught pupae fr om North West Frontie r Province, Pakistan. Mosq. News. 39: 756-762. Suzuki, A., Y. Tsuda, M. Takagi, and Y. Wada. 1993. Seasonal observation on some population attributes of Aedes albopictus females in Nagasaki, Japan, with emphasis on the relation between body size a nd survival. Trop. Med. 35: 31-99. Telang, A., and M. A. Wells. 2004. The effect of larval and adult nutrition on successful autogenous egg production by a mosquito. J. Insect Physiol. 50: 677-685. Telang, A., N. A. Buck, and D. E. Wheeler. 2002. Response of storage protein levels to variation in dietary pr otein levels. J. Insect Physiol. 48: 1021-1029.

PAGE 109

109 Timmermann, S. E., and H. Briegel. 1993 Water depth and larval de nsity affect development and accumulation of reserves in laboratory popu lations of mosquitoes. Bull. Soc. Vector Ecol. 18: 174-187. Timmermann, S. E., and H. Briegel. 1999. Larval growth and biosynthesis of reserves in mosquitoes. J. Insect Physiol. 45: 461-470. Turell, M. J., M. R. Sardelis, D. J. Dohm, and M. L. O'Guinn. 2001. Potential North American Vectors of West Nile Vi rus. Ann. NY Acad. Sci. 951: 317-324. van Riper III, C., S. G. van Riper, M. L. Goff, and M. Laird. 1986. The epizootiology and ecological significance of malaria in Hawa iian land birds. Ecol. Monograph. 56: 327-344. Villavaso, E. J., and C. D. Steelman. 1970. Laboratory and field studie s of the Southern House Mosquito, Culex pipiens quinquefasciatus Say, infected with the dog heartworm, Dirofilaria immitis (Leidy), in Louisiana. J. Med. Entomol. 7: 471-476. Vinogradova, E. B. 2000. The Culex pipiens complex, pp. 4-45. In E. B. Vinagradova [ed.], Culex pipiens pipiens mosquitoes: taxonomy, distri bution, ecology, physiology, genetics, applied importance and control. Pensoft Publishers, Sofia. Walter, N. M., and C. S. Hacker. 1974. Variation in life table characteristics among three geographic strains of Culex pipiens quinquefasciatus J. Med. Entomol. 11: 541-550. Williams, G. C. 1957. Pleiotropy, natural selection, and the evolution of senescence. Evolution 11: 398-411. Xue, R. D., J. D. Edman, and T. W. Scott. 1995. Age and body size effects on blood meal size and multiple blood feeding by Aedes aegypti (Diptera: Culicidae). J. Med. Entomol. 32: 471-474.

PAGE 110

110 BIOGRAPHICAL SKETCH Sean McCann has had a passion for the natura l world ever since he was a young child. He graduated from the University of Victoria Biology Honours program in 2004, studying with Dr. Richard Ring and Dr. Neville Winchester. Hi s interests include photography, the outdoors and animals of all sorts.