HOST MI~CROFLORA RELATIONSHIP~ OF
VECTORS OF CANINE HEARTWIORM DISEASE
Dale R. Hamnilton
A Dissertation Presented to the G-lradute Councl o
the University of Florida in Partial: Ful-fillmicgn of
the Requirem~ents -for the Degree of DC~~tor of P~i losophiy
Univer-si iy of Florida
Dale R. Hlamilton
dedicatedl ci my mrke:-ir and dasd,
whose patien~ce hiad nzo i~mitn,
furo their' son iJho -Is le-ss for~tunLte.
I am proud to embark upon my scientific career, distinguished by
the influence of Dr. R.E. Bradley. Dr. Bradley has my gratitude and
my admiration for having brought me to this moment in my professional
I am grateful to Dr. H.L. Cromroy, Chairman of by Supervisory
Committee, who has exemplified what to me are the best qualities of
scholarship and teaching. Dr. J.F. Butler deserves my lasting thanks.
His door and his mind were always open.
Dr. Donald Weidhaas generously opened the door to the USDA Agri-
cultural Research Service, Insects Affecting M4an Laboratorty, of which
he is its director. I appreciate the laboratory facilities, office
space, and most of all, the vast resources of experience made available
Dr. Jai K;. Nayar of the Florida Medical Ent;om;ological Laboratory
took time from his own research to recount his invaluable experience
with gnotobiotic insect culture.
I am~ obliged to my friend and lab technician, Louis Ergle, who
tried in vain to impart the virtues of laboratory cleanliness. Barbara
Hudgens engaged in a heroic one-wdoman struggle to interpret the tangled
sk~ein kn-own as "a first draft".
I ami also grateful for my place at the Red Lion bar where I re2-
ceived much valuable counsel throughout myv tenure of study.
Finally, a special thanks to Ginny. To her, I dedicate all mly efforts
This investigation was supported by NIH training grant number 5 TO1
AIOO 383-04Sa from the Institute of Allergy and Infectious Diseases.
TABLE OF CONTENTIS
A.CKNOWLEDGEM4ENTS .................. ... .. ............ iv
LIST OF TABLES .............................. ........... Vi
LIST OF FIGURES ....................... ... ........ .......... vii
ABSTRACT ................ .................. .................. viii
INTRODUCTION .. .............. ......... ............ ..... 1
LITERATURE REVIEW ................ .. . ..... ..... ......... 10
Dual etiology of parasitic disease ................... ...... 24
Microflora associated writh mosquitoes ................... ....26
Statement of the problems ............ .... ............... 37
RESUILTS ..................... 39
I. Bioassy of Early Vector Miortali-ty Following
Dirofita~r~ia immzitici Leidy Infection ..................** 39
II. An Integrated System for the Production of
Gnotabiotic Anopheles qua~drifmacu~latus Say ............ 47
III. The N~ative and Midgut Bacterial Flora of Several
Wild and Colonized Mosquito Spcies ....... ............ 60
IV. Effects of Bacteriological Flora on the Ear~ly
Development of Divofilaria immi~tic Leidy in
Alnopheles qua~drirm,cula,~tus Say .......... ... ......... 89
DISCUISSION AND CONCLUSIONdS ........... .... ... ......__. '105
BIBLIOGRAPHY ...., ............. . .. ._ ............,113
BIOGRAPHICAL SK:ETCH .............. ..........................13
LIST OF TABLES
1-1. Effects of Diff:erential Filaremia on
Mortality of An~opheles qjuadrim~acouZatus3 Say .............. 42
I-2. Effects of Feed-ing Dead Intact, and Homogenized
Microfilaria to Femnale Anropheles3 quadrimaculatus Say .... 43
I-3. Effects of Feeding Homogenized Body Parts from
Females Fed Infective Meals 24 Hours Previously ......... 44
II-1. Composition of Diet for Rearing An~opheless
quadr~imaLcuilaus Say ................... ...............~... 55
III-1. Species Com~position of Midgut Bacterial Flora
in 5 Groups of Mosquitoes ............................... 66
III-2. Correlation Coefficients of the Frequency of
Occurrence of M~idgut Bacterial Species .............----- 82
LIST OF FIGURES
II-1. Gnotobiotic Arthropod M~odule (GAM) Used for the
Maintenance of Blood-Feeding Aniophele c quad?rfrimaculatu/s
Say .................................... ........ 52
II-2. A Battery of 4 GA~s in Position for Blood-Feeding
in Heating M4anifold ....................................... 54
III-1. Procedures in the Presumptive Determination of
Bacterial Genera ......................................... 65
IV-1. Type I-Ib Larva in Malpighian Tubule "Squash" from a
Conventional Nlosquito 48 Hours After Taking an
Infective Blood Meal ................... .................. 95
IV-2. Typically Unchanged Type I Larva Recovered After 48
Hours in a Mosquito Mionocontaminated with Salm~onella sp. ... 96
IV-3. Early Type II Larva Recovered After 48 Hours in a
Mosquito Monocontaminated with Acinetobacter calcoacetious.. 97
IV-4. Type II Larvae Recovered After 48 Hours in a Mosquito
Mlonocontarninated with BacilZurs cereus ................... ... 98
IV-5. Early Type III Larvae Recovered After 48 Hours in a
Mosquito Mionocontaminated with LactobaciZllue sp. .......... 99
IV-6. An Encapsulated Type I Larva Recovered After 48 Hours
in a Mosquito Monocontaminated with Non-pathogenic
coryneb5acterl:um sp. .......................................10
IV-7. Typical "Sausage" Stage Type IV Larva Recovered After
48 Hours in a Gnotobiotic Mosquito ........................101
IV-8. Tw~o "Sausage" Stage Type IV Larvae Hlithin a Tubule of a
Gnotobiotic Mosquito .......... ......... ..................102
Abstract of Dissertation Presented to the
Graduate Council of the University of Florida
in Partial Fulfillment of the Requiremnents
for the Degree of Doctor of Philosophy
HOST MICROFL.ORA RELATIONSHIP OF
VECTORS OF CAN~IIE HEARTWORMV DISEASE
Dale R. Hamliltton
Chairman: H.L. Cromroy
Co- cha irma n: R.E. Bradl ey
Major department: Entomology and Nematology
The phlenomnenon of early vector mortality of Dir~oji~iLaric. imitici
L~eidy infected mosquitoes wJas investigated by means of a bioassay -tech-
nique. Anoplheles quadrima,~3cuZabuss Say w~ere fed differentially filaremic
blood meals, living and dead microfilariae, and special adjuncts such
as homogeniz~ed mosquito body fractions. At least 25"! mortality wajs
produced r inmosquitoes ini the first 3 days following ingestion of living
miicrofila riae. Nlei their d-ad, intact, nor- homogenized larvae, whole
mosqui to bodies, nor body fractions refproduced this miortal ity, suggesting
that ear-ly vector death is a function of the kinetics of living micro-
An integrated systemn of diet, technique, and hardwairee was developed
for the productionl of gno0tabiotic blond--feeding arthropods. This system
produced the fir-st successful rearing of An,. qua"dr~iiimr;L7atulaba under
A quantitative and qualitative bacteriologic evaluation was
accomplished on the midguts of several hundred adult female specimens
of colonized in. qu~adrjimacuiZatu-s, wild An. quladri-macuilatus, A-nopheles
crurcian-s Weidemanrl Anophel~ec cruciansr complex, and Aedea ii~jrnfcirmt
Dyar and K~nab.
The total midgut bacterial count, and the species composition of
midgyut flora was distinctive for each group of mosquitoes. The most
frequently isolated bacterial species for each group of mosquitoes were:
A. inir.-zmatis, non-pathogenic Coirynebacteri.:m sp., wild An. qu~adrimaculAiatus3,
Bnterobacte~r cloacae! (Jordan), and Ac~inetobater calcoace~t~ions (Beijerinck),
colonized Ani. quaodrimaev7Yat lue Bacill:L20 ctereuo Franklanld and Frankland,
An. cruciiians, Allealigenzes ~faeccali~s Castellani and Chalmers, and An.
cr~-ucian complex, Sa~n0trone!Za sp. Data analysis revealed several signi-
ficant (P .001) positive and negative correlations among the various
bacterial species comnpris-ing thle normal flora of these 5 mosquito groups.
A proportion of~ bacteriologically sterile specimens were encountered
in 4 of the 5 mosquito groups. Sixty-one per-cent of w;ild An. quaclh~rimcun-
latvos were sterile while 410% of the colonized An. qua~d2ima7culati.u s wer'e
sterile. An. crutciann and An. crucians complex were 36% and 43;: sterile.
All specilenss of Ai. infiine alz~ examined w~ere bacteriologically sterile.
Pure cultures of bacteria isolated froml these mnosquito-s wJere used
a s mnon ocontami;i na7nts of gjno tobi oti c ;ln. quadi~macula;7lln tus. Thiese mosqui toes
as well as conventional and gnotobiotic mosquitoes were subsequently: in-
fected with D?. immithii infective larvae. Forty-eight hours post-
i nfecti on all mosqui toes w~ere ki lled. The degree of filarial develop-
mnent w~as recorded by light and phase photomicroscopy, and typed ac-
cording to their gross dimensions.
The early development of D. irmmi~tis larvae proceed independently
of bacteria or bacterial mediation in the midgut. Sausage stage larvae
appeared within the first 48 hours in gnotobiotic mosquitoes, but re-
quired at least 72 hours in conventional mosquitoes. Nematode de-
velopment appeared to be marginally improved in the complete absence
of bacteria and differentially retarded in the presence of pure cultures
of certain bacterial species. Pure cultures of LactoaSiZelvjs sp. and
B. cerceui, were found associated with more advanced larvae, while SaZ-
monella sp., Al. calcacetiouls and Corynebacterium~hn sp. monocontarminarnts
were found in the presence of less developed larva. If the presence of
midgut bacterial flora is inversely related to the development of D.
immritic in the mosquito, it is suggested that the "naturally" sterile
specimens in the field are the most effective vector of D. immritis.
Caninle heartwor-m disease research in the last decade has evolved
into tw~o distinct efforts. The first approach has been towlards develop-
ment of adequate chemotherapy to replace the potentially hazardous
arsenicals presently in use. Heartw~orm prophylaxis, although effective,
involves considerable risk as well as bothersome clinical management.
Development of a vaccine or imrmunizing agent against canine heartw~orm
disease, although promising (Otto, 1970), is not imminent and miust aw~ait
fuller understanding of the mechan-ism of imm;unity to filarial infection.
The second approach, and historically the miost suiccessful, is
direccted against the arthropod vector. Di?ofitar~ia fronit Le'idy, the
etiological agent of canine heartworm disease, depends upon passage
through a mosquito to th-e infective stage. The traditional methods of
controlling such arthropod borne diseases such as filar-iasis, yellow fever,
malaria, and trypanosomiasis have been to release sufficient pesticide
into the endanger-ed area. This has been justifiable, because it is most
certainly DDT and its descendants that hvav done the most to reduce, and
in places eradicate, malaria, typhus, filariasis, trypanosomiasis, and
many other of the world'ss worst killer diseases (Provest, 1972).
Inl recent years, it has become dramatically apparent that the
traditional netihod of vector control using insecticides is no longer
effective. The Universitiy of Califor-nia's Mosquito Control laboratory
recently reported that the common pasture mosquito Asda!! nig:i.-naculia;
(Ludlow) is now~ resistant to every available insecticide (Kendrick, 1970).
Boardman (1973) quoted a report of the Agricultural Experiment Station
at the University of California which reported the fact that the two miost
important species of mosquitoes in California had developed near total in-
secticide resistance. Elsewhere, the use of organic insecticides against
mosquito larvae has led to so mluch resistance, in Cuter f~atigans Wi edemann
especially (Brown, 1967), that control of filariasis and several arbo-
viruses is now one of the greatest problems of vector control. According
to a recent review (Busvine and Pal, 1969), resistance to DDT, BH!C, and
dieldrin is severely affecting control, in one place or another, of malaria,
yellow fever, filariasis, typhus, and plague, while certain other vector-
borne diseases have so far not been affected by pesticide resistance, e.g.,
onchoceriasis, Chagas disease, typanosmiasis, and le-ishmnaniasis.
To compound the problem, killing mosquitoes with insecticides is
becoming prohibitively expensive. The logic of pest control demands more
specific chemicals. However, the more specific a compound is the more
limited its use, and the more costly it is to produce. A decade ago, it
cost less than half a million dollars to market a new~ insecticide. The
same development job in 1970 cost from 3 to 5 million dollars (Kendrick,
1970), and today even more. At a time~ when mosquito abatement districts
are demand-ing new insecticides to replace those of declining value,
manufacturers have less incentive to produce then and particularly the
narrow~ spectriml m;aterialls designed for limited uses.
The persistence of pesticides in thle environment, now thoroughly
documented (Woodwell, 1970; Mriller and Berg, 1969), has led to increasingly
miore stringent regulations to govern their uses aInd appl-ication. The
Federal Environmental Pesticide Control Act of 1972 for example, established
the Environmental Protection Agency which promptly banned the use of DDT.
The net result of increasing insecticidal resistance, obsolete.
pesticides, and legislative restraints is that mcosquitoes are increasing
in number, and w~ith themi, canine hefar-tworm disease. Results of a nation-
wide geographical survey of canine filariasis taken in 1955, have made it
abundantly clear that canine heartworm disease is expanding its range
(Young, 1955). The response of 2,337 veterinary practitioners throughout
the United States indicated that the distribution of heartworm was primarily
in the states bordering the Gulf of Me~xico and the Atlantic Ocean, but ex-
tending northward to illinois and central lowa (Young, 1955). Marquardt and
fabian (1966) collected reports of D. iirnwilino in Florida, Georgia, Alabama,
Louisiana, Mississippi, Ne~w Jersey, Tennessee, Maryland, Arkansas, Pennsyl-
vania, Rhode Island, and Hawaii. Subsequently, enzootic dirofilariasis has
been reported in Indiana, Michigan, Ohio, Canada (Otto and Buaman, 1959),
Mininesota (Schlotthoauer and Griffiths, 19)64) and Connecticut (Hirth et al.,
1966). In a 1973 statewide survey of Florida, Bradley* reported that 27
north Flor-ida veterinarians diagnosed 4,648 cases of caninie heartworm in
1973, an increase of 22% over the previous year. Seventy-seven central
Flor-ida veterinlarians and 84 south Florida practitioners reported 23,212
and 8,658 cases, respectively, 1973, an increase of more than 50% over
the 1972 case level. Eighteen west Florida veterinarians diagnosed 3,927
cases in 1973 which consti-tuted a 33% increase for the year.
Nlot all mosquito species, however, are capable o-f transmitting
;. ?IIII.cibS. Demirick and Sandholn (1966) list 50 mosquito species known to
be successful vectors in the United Statefs. Ma~cnonald (1971) has compiled
a list of 107 species in 6 genera of Culicidae which support the development
of filarial larvae. Ludlam and Jachlowsk~i (1970) report 64 species of
mo~squitoes in 5 genera in wihichh comcplete larval development of s. immi~si
* R.E. Bradley, Personal Com~munication, 1973
can take place. Yen (1938) described differences in relative susceptibility
to infection in 12 mosquito species. Travis (1947) demnons-trated the relative
efficiency of 6 species of mosquitoes from Guam, and Intermill and Frederick
(1970) reported a similar comparison of Ryruku Island species. Kartman (1952)
also reported a quantitative evaluation of the developmentt of D. immiiis in
7 common mosquito species. Furthermore, evidence has accumulated that indi-
cates that the same species of mosquito may, exhibit varying susceptibility
to D. imnit~ic in different laboratories. K~artmran's (1950) monumental survey
of the world literature from 1901 to 1949 indicated that the experiments of
over 30 workers show that Ac~des aegypti (Linnaeus) varied in susceptibility
to Wutchrereia bannerofti (Cobbold) from completely inhibiting the! parasites'
development of allowing maturation of infective larvae.
Eveni within successful species and refractory species, there are
individuals who show varying susceptibility to infection with D. inuritis.
A recent epizootiologic study carried ouit by Bradley (1971), surveyed Hills-
borough County, Florida, an area heavily populated by known vectors of
D. iinvitis. Only small numbers of parasites were found in resident dogs,
which Bradley (1971), concluded may indicate some factor being exerted on
the vector population to limit the exposure of dogs to infective larvae.
Miccreevy et al. (1970) in a survey of 5 counties in northern California
failed to find a single case of D. im~mitis inl 515 pound dogs. The same
authors surveyed the clinical records of dogs examined for heartworm at the
University of California's School of Veterinary Medicine from 1957-1968.
Only 13 cases of D. irnvitifi infection could be con-firmied. Of the known
species of mosquito vectors, 9 species werne reported to occur in Calif~ornia,
and 6 were reported in counties where D. i!mmais was known to occur
(M-c~reevy et al., 1970). A survey of 50 wild foxes trapped within the
geographical limits of a hyperenzootic area in Georgia proved negat-ive for
D. 7jwnitis microfilariae (Walton et al., 1963).
Regional differences in relative susceptibility to infection had
also been noted. R~oubaud (1937) reported that Assam and Tanganyika strains
of Ai. aegypti permitted normal development of D,. wnimmics buit Cuban strains
infected from the same host were entirely refractory. Newton, W~right and
Pratt (1945) reported that Puerto Rican strains of Peorophioia con~finiz had
a 12% infectivity potential for W. bancroffi, whereas United States strains
had an 80% rate.
The basis of vector susceptibility fo filarial larva has been hypothe-
sized since Noe (1908) first reported the destruction of microfilariae of
Dipebtnonzema grransi N~o~e, in thie gut of its tick host during the latter
stages of feeding. Rozeboom (1973) had observed that filariae exhibit a
strict vertebrate host specificity, but the insect vectors for a given para-
site may include many species among two or more genera. These parasite-
vector associations, he noted, have arisen as the parasites evolved and ex-
panded their geographic distribution and have depended less upon the taxo-
nomic position of the insect species than upon the fortuitous existence of
genetically controlled factors for susceptibility. WJeinstein (1973) has
made the following generalizations regarding the specificity of development
of filariae in arthropod hosts.
1. There is generally a higher degree of specificity in the
vertebrate phase of the cycle than in the invertebrate.
2. Most recognized vectors are mosquitoes.
3. Filarial development in the arthropod takes place in only
one of 3 types of cells (Mialpighian tubule, thoracic flight
muscle, and fat body), and the specificity of cell type is
virtually abDsoute fur a particular species of filariid.
4. Development is intracrellular.
A genetic explanation for susceptibility of invertebrate hosts to
certain parasites has received substantial support. Huff (1941) determ-ined
that the susceptibili'ty of cular piipices LinlnaEu( s to P~oameld iumn cazthemerrium
Hartman is a recessive character due to a single pair of genes and in-
heritable along the lines of thle classical M~endelian 3:1 ratio. Trager
(1942) produced a strain of A. aegypti mrore susceptible to infection wi th
PZasmodiium Zophurae Coggeshall than the original colony strain. A report by
the R~ockefeller Foundation (1948) indicated that it w~as possible to raise
the susceptibility of Anopheles quLad~r-imacu7Y atus Say to Pla::mod-iu galiamesw n
Brumpt from 20% to 100%. By selective breeding Nicks (1949) elevated the
susceptibility of c. pipien~s to PZa;emodCium elongatium Huff and Bloomr from 13%/
to 49% within 6 generations.
Evidence to the contrary ha:: been advanced by Boyd and Russell (1943)
who could obtain no clearly defined results in attempting to select a strain
of Anz. quadrimnaculatus more susceptible to Plasmodiurm vivrw: (G:-assi and
Feletti). Jeffrey (1949) similarly failed to produce a strain of Anophelea
albopictus (Skuse) more susceptible to infection- withC P. Zophuroae and
Horvanitz (1947) could obtain no genetic effects in 6 generations of A.
aegypti selected for susceptibility to P. ga2%iincea~.
A genetic basis of mosquito susceptibili ty to filarial nematodes was
first suggested by Roubaud it al. (19353) and Roubaud (1937) w!ho noted
differences in the susceptibility rate between colonies of AI. aegy~ptC
originating from different geographical areas and exposed to the same strain
of D. imanitis. Ramachandran et al. (1960) reported similar variations in
susceptibility of Al. aegyp~ti for^ BrG1h- m77,cai (Brug). The response of a
colony of mosquitoes to artificial selection pressure for changed suscepti-
bilit~y was noted by Kartman (19533i)wi-th D. imnitics and by Thomias and
Ramiachandran (1970) for Wuzche~rcria~ sp. in: DAlrx pipit'" ns frabigno Wiedman.
N~acDonald (1962a, 1962b, 1963a, 1963b) determined that susceptibility
to infection of A. aegyipti w~ith sub-periodic B. malayi was controlled by
a sex-linked, recessive gene called FM which exhibited incomplete dominance.
Subsequently, MacDonald and Ramachandran (1955) reported that this gene
controlled development of 5 str~ainls of Wu~chereviai sp. and DvGi~:a sp. in
the thoracic muscles of Al. aegypbi, but apparently had no effect on the
development of D. inmitic or Divoufitarila repensc Rail let and Henry in~ the
Malpighian tubules. McGreevy (1972) finally demonstrated that susceptibility
of A. aegypti to D. jmmitis was also controlled by a sex-linked, recessive
gene and that penetrance of this genie was incomplete. Moreover, the gene
that controls the development of Dn. inside~i in the M~alpighian tu!bules is
not identical to the FM gene which controls the development of Wuichercria sp.
and Brugia sp. in the thoracic muscles. M~acDonald (1973) cited unpublished
woirk by Oblamine and MacDonald that indicates the susceptibility of c. p.
fatigans to D. pahzangii was also controlled by a sex-linked, recessive gene.
Kartman (1952) considers that genetic factors may operate independently
of the phylogenetic position of the host. The efficiency of 2 closely re-
lated mosquitoes as vectors for D. imm-itis is not necessarily comparable.
Instead, susceptibility seems to be an inherent characteristic of the in-
dividual mosquito. Hu (1937) gave C. p. pallena double spaced infectious
feedings of W!. !:aw;rofi' in an effort to determine whe!ther susceptiblity
to infection wJith filaria in an inherent characteris-tic of the individual.
Hie reported that some2 of the fem-ales were susceptible to the first but not
to the second infective meal, anid vice versa. Similarly, imm~unity to re-
infection w~ith D. i:mitis would not develop ill Ausph :-,Zes I;uncT;ti enni: (Say),
Aedes cineru~s Mleigen, and ZedeZn: eviiseratus Say (Phlillips, 1935). Ka~tma-n
(1953) noted an unusually high susceptibility for D,. i;-r.itic drmong certain
individuals of Haw~aiian str-ain of A. cogyp'ti, a species generally refractory
to ths filriae.Furthermore, individual respione of a mosquiohoti
distinctly specific to the species of parasite. Kartma7n (1952) could
demonstrate no correlation between sulsceptibility of A. aegypti to mixed
infections of D. i:~nvitic and FoZeye~ll sp. Each species of filaria reached
its normal location within the mosquito without interference from the other
species. This is consistent with the observations of Huff (1930) on in-
fection of C. pipienc with several species of avian plasmodia, and with
those of Moorthy (1938) on the development of Canattanlusu sp. in cyclops
previously infected with D~-racen-Lutus sp.
A more critical examina-tion of the individual mosquito, rather than
a species as a unit, seems to be warranted. The factors and circumstances
which serve to distinguish the -individual mosquito are manifold, and may
have an important effect on the individual's susceptibility to ma~ny parasitic
agents. One of these distinguishing factors may be quantitative and quali-
tative differences in the composition of intestinal microflora.
Mosquito species tend to select very specific nesting sites and
subsequent rearing sites for the larvae. Considerable diversity exists
between species in their choice of sites which range from vernal ponds
(Wills and Fish, 1973), pools and flood w~at:ers (Jamezs and Harwood, 1969)
to tree holes and automobile tires (W~omack, 1971), and inl certain da-ter
holding plants suich as bromeliads, heloconids and pitcher plants (Gillette,
1972). The bacterial and fungal populations of these selected microhabitats
reflect the diversity of their environment.
Newly hatched mosquito larvae are indiscriminant filter-feeders which
voraciously invest anything of a specific particle size. Since this
particle size is usually less th~an 100 microns, bacteria make up a sub-
stantial proportion of thle larval diet. This fortuitous acquisition of
maicroflora apparently established the individual's lifetime comnplemlent
of internal bacteria, which is both pspcies specific and individually
distinct. Quali tative and quantitative differences in the composition
of this microflora may represent one of the factors mediating vector
susceptibility to filarial infection. The passage of filarial larvae
into and through the gut of suiscep~tible mosquitoes evidently r-equires a
critical set of conditions since attempts to culture infective filaria
larvae have me-t withh little success (Weinstein, 1970, 1972; Rothstein and
Brown, 1960; Sawyer and W~einstein, 1933.4, 1963b; Taylor, 1960b;Wood and
Suitor, 1966; Cupp, 1972). Residen-t microflora may indirectly interact
with the infective larvae, or merely mediate an intestinal climate
favorable to survival of the parasite.
The demonstration of an interaction between larvae of D,. i~mmtis
and resident bacteria, may provide a future means of biocontrol for both
anthropod and niematode.
Subsequent to feeding on a filaramic dog, 3 stages in the development
of D. inmniti8 occur within the mosquito (Taylor, 19360). A moul-t marks the
onset of each stage terminating in the third stage infective larva. M~icro-
filariae taken in with the blood remain in the stomach of the mosquito for
the first 24 hours. During the next 24 hours they migrate to the Malpighian
tubules. Further development occurs at thie distal end of the tubules until
the fifteenth or sixteenth day. The larvae spent the first 6 or 7 days in-
side the cells of the M~alpighian -tubules. Suguri et al. (1969) have de-
scribed this zone on the Malpighian tubules as rich in mitochondria and
course endoplasmic reticulum. The brush border has a multitude of vertically
arranged protoplasmic processes each containing a mitochondria whose cristae
run parallel to the process. Fromn thle ninth to the fifteenth day, larvae
are found in the lumen of the tubules. By the fifteenth day the larvae are
ready to break out of the tubules and migrate through the mosquito, to occupy
the cephalic spaces of the head and proboscis (Taylor, 1960). Transmissionn
of infective larvae occuISrs with the next blood meal. Al-though many arthro-
polds have been evaluated as vectors of D. ZinwZith,, only mnosquitoes appear to
be entirely competent (Ludham and! Jachowski, 1970).
Kartman (1952) observed that althoulgh microfilariae of D. C-invitic are
knownl to be capable of metamonrphosis in many spcies of m~osquito-s, there is
a dist-inct pauicity of host species in whic-ih it maiy adequately reach the in-
fective larval stage. The same writer introduced thec concept of "host
efficiency," a ratio between theoretical numbers of microfilariae ingested
by a group of mosquitoes and the total number of developing larvae found in
them during a specified period, and "infective potential," a similar ratio
involving only infective larvae. Kar~tmian (1952) reported that the host
efficiency of Anl. qua~dritaimculatu:; was approximately 20 times better than
either Al. aegypti or Culex qutnquefus-rc auSnllr Say; and that the infective
po-tential of An. quzadr:imacuZabus was 50 times greater than A. ae~gypti and
20 times more than C:. quinquefansciatus. Infection of C:. pipiene and C.
quinquefas~!ciatus showed that both the hiost efficiency and the infective po-
tential of C. quvinquef,~aceiatus~. were 2-1/2 times greater than those for C.
pipiens. Ewart (1965) correlated variations in vector potential of CuZe-,
Aln3,he~ine, and lede-s mosquitoes to failure of Brugia. pahangi Buckley micro-
filaria to exsheath. Weatherby et al. (1971) reported a similar case of host
efficiency in susceptibility to P. galinazcewnm by feeding mosquitoes whole
body extracts of both susceptible and refractory mosquitoes. CuiZez pipiens
extract (refractory) fed to susceptible A. aegypti reduced the number of
oocysts developing by 50% while concentrations of 75%etatreue h
numbers to only 34% of the number in the controls. Clearly, some factor
present within refractory mosquitoes, and transferable,w~as capable of reducing
the host efficiency of susceptible mosquitoes.
A problem fundamental to the production of experimental Di. im~ibia in-
fections has beeni survival of the mosquito -following uptake of m~icrofilariae
fromn infected caninle blood. Various authors including Yen (1938), Kershaw
ot al. (1953), Bradley (1953a, b), Hiusain and Kershaw (1956), Wefbber and
HawJking (1953), Pistey (1959), Symes (1960) and Weiner and Dr-adley (1970)
have reported vctor- mortality approaching7 1003 in the first fewi days follow-
ing the blood m~eal. A common antecedenti to death appears to be distenision of
the abdomen persistinig 2 to 3 days, whiich Weiner and Bradley (1970) interpret
as a result of the improper digestion of blood. Additionally, it appears
that vector mortality increases proportionately with the number of micro-
filariae ingested (Rosen, 1954; Kershaw et al., 1955; Wlharton, 1957; Kartman,
1952). Kartman (1952) suggested that variations in vector survival w~ere
sometimes dependent upon genetic strains of the same species, and later
(19534)observed a direct relationship between numbers of microfilariae fed
to mosquitoes and vector survival. Diet (LeCorroler, 19J57), environmental
temperature and humlidity~J (Galliard, 1957), but not atmnospher-ic pressure
(Williams, 1959) have been shown to affect survival of filarial-infected
m~osquitoes. Laurence (1963) had estimated the natural mortality of 2 infected
mosquito populations in the field. c. fa;tigans and An~opihelrs pediitae~nia~tus
(Leichenster) in south Inidia had a daily miortality of from 14% to 24% during
a season favorable for survival.
Kershaw et al. (1953) investigated the effect of microfilaria of D.
immjitia~ on the longevity of A. aegypti under controlled laboratory conditions.
Infected mosquitoes had a constant but very low survival rate during the
first 5 days corresponding to the migration of microfilaria from the midgut
to the Malpighian tubules of the mosquito: a constant and high survival
rate from the fifth to the twentieth day, corresponding to the development
in the tubules; anid a very low but constant survival ra7ie flrom the 21st to
the 25th day, corresponding to the presence of infective larvae in the hemro-
coel and in the head and! mouthpartss of the mosquito.
Septicemia induced by emigrating microfilarid may be another contribution
to vector mortality in the first few dayvs -following infection. Intestinal
miicroflora, fortu-itously acquired in the larval stage, ar~e unable to multiply
in the conditions prevailing in the gut of healthy insects (Bucher, 1960).
Substantial mortality from bacterial infections in laboratory colonies had
been reported w~hen environmental conditions favor thel multiplication of one
or more of the bacterial species already present inl the guit (Prinsloo, 1967).
Chu and Toumanoff (1965),orepote that most of the bacteria isolated from
laboratory colonies are Gram-negative which K~alucy and Daniel (1972) concluded
are "potential pathogens" which require predisposing factors such as damage
to the cuticle or gut wall in order to penetrate the hemocoel. The bacteremia
resulting from multiplication in the hemocoel caused death of the host.
Nematode-intro3duced septic conditions are not unknown in the literature.
Most of the infective juveniles of the DD-136 strain of Necapteebana car~poca-
sae W~eiser contain cells of a specific bacterium, identified and described as
Alchromobacter -nemato;/h~ilis in the ventricular portion of their intestinal
lumen (Poinar, 1972). Soon after the nem~atode reaches the hemocoel of a para-
sitized insect, such as G02Zleria larvae, the bacterial cells are passed fromn
the intestine out through the anuis into the hemolymph. The bacteria then
rapidly multiply in the hosts' body causing a fatal septicemia. In mos t in-
stances it is the bac-teria that actually kills the host-usually wiithin 48
hours; however, most nematodes without their associated bacteria are also
able to kill the host within a short period. Without the Achro-?oba~ter
bacteria Poinar (1972) reports, N;. carp~oozpnea is unable to gain access into
the insect hemriocoel, thus the relationship between A. LneatoyhZi lic and
N,. carpocapa~ce m~ust be considered mu~talistic (Poinar and Thiomas, 1967). Such
bacteria ar-e undoubtsd ly as:sociated w~it-h all of the 11 knlownr species of
WNyop~chanazi (Wdelch:, 1965).
Large numbers of m~icrofi'lariae ingested wi th blood iieal have been
shown to interfere wi, th the formation of thle p-tritrophic mem;brane with
fatal r-esuilts (Lew~is, 1953). Wadnson (1950) noted that the blood contained
in thie stomach of Sier2Zin ;um downocu Th-obald killed by heavy m~icrofilariae
infection, appeared to be undigested. Similar observations w~ere recorded
by Roubaud et al. (1938) and Mlackecrrar. (1953) on other filar-iae-host com-
binations. Lavo-ipierre (1958) regarded the mos-t likle~ly cause of death in
heavily infected arthropods as "indigestion" due to extensive damage to the
peritrophic membrane rendering thle digestion of blood almost impossible.
The same author found it impossible to demonstrate the peritrophic membrane
in dying heavily filaremic A. aeSYpti with und-igested blood meals, whereas
thle membrane wdas present in the controls in which the blood meal was in an
advanced state of digestion. Bain and Philippon (1970) found the peritrophic
membrane intact in AniOpheles3 stdphnC~1i Liston and S. dan':osum; given a blood
meal containing reptile microfilariae. Furthermore, they concluded that the
peri-trophic membrane did not prevent the crossing microfilariae from the
blood meal to the midgut. Lewis (1953) reported that peritrophic membrane
formation in S. dawnmocumn following the blood meal was critically important
to thle survival of onzchocerieavolvuilus Leuchart. Only the filariae which re-
main in the anterior portion of the midgut are able to pass backward (after
formation of the membrane) into the space between the membrane and the gut
wall in order to pierce the stretched waills of the distended stomach.
Wigglesworth (1965) described two different methods of peritrophic
membhrane formation around the food bolus. The membrane may be delaminated
from most or all of the midgut epithelial cells, or it may be secreted from
specialized cells in the anterior mnidgut. The mridgut opithelium of mouse-
fed A. asegypti produces a granular secretion for up to 15 hours after feed-
ing and the peritrophic mlembranee is apparently formed from this (Bertr~am
anrd B-ird, 1961). The membrane formed by the first method generally acquires
the shape of the food, bo'lus, while that formed by the second method is
usually cylindrical. In larval mosquitoes, the membrane is formed by the
second method (Wigglesworth, 1930; Richards and Richards, 1971). The mem-
brane in adult mosquitoes is formed by the fiirst method (Freyvogel and Staubli,
1965; Freyvogel and Jacquet, 1965; Stcher, 1957). Romoser andr Rothman (1973)
published the first report of the presence of peritrophic mem~branee around
the meconium of a pupal mosquito. Romoser (1974) subsequently demonstrated
a second membrane, closely applied to the first which is formed somewhat
later in the pupal period. Howard (1962) reported that the peritrophic mem-
brane appears about 12 hours after At. aegypt-i feeds, and increases in size
up until 36 hours, and becomes brit-tle by 48 hours. As the amount of blood
in the midgut decreases, the peritrophic membrane is fragmented by the con-
traction of the midgut muscles.
Zhuzhilov (1962) reported that the peritrophic membrane was apparent
in A. aegypti within 20 minutes of taking a blood meal. In contrast, Frey-
vogel and Staubli (1965) could not discern the membrane in this species
before 8 hours after -feeding. Richardson and Ram~oser (1972) reported that
the membhrane was visible in Ai. twiseriauer 20 minutes after taking the
Stohler (1957) noted that the memabrane adhieres to the blood parcel,
not to the midgut cells. Although then membrane in Aeder; sp. passes through
stages described by Freyvogel and Staubli (1965) as viscous, elastic, solid,
and finally fragile, the membrane in Anroph-?en sp. never develops beyond a
fragile membranes. The same authors also report::d that 3 species (A. cofuPi;~
rinopho~leS sawh71i*., Gilals, Ani ScOIphem:1 i) whilch no-nmally formn a pe~ritrophic
nemb~-rane do nlot do so comnpletely if the2y ingesjt onlly a small quan:t~ity of
blood. Richairds and Richards (1971) found n~o tradce of a m~embranez in adult
male or unfed adult -female A;. aaegyphi 1 to 3 days post emergence. Older
adult fem~ales given a blood meal produce a peritrophic mem~brzne from material
secreted by a ring of midgut cells in the anterior half of the proventricular
The rate of destruction miicrofilar-iae during the first 48 hours after
ingestion determines vector mortality to a considerable extent. Hughes (1950)
observed that miicrofilaria of rLitomociozdes; ingested by or~nithon;iuccuIs ba~coti
(Hirst) mites w~ere phagocytosized by certain cells in the midgut. Kartman
(1953) regarded the initial destruction of microfilariae in the midgut as a
process of digestion, although, as he p~ointed out, death may not be caused
by digestive enzymes but rather by other factors ini the midgut or salivar~y
secretions. Digestive destruction of inappropriate parasites is an early
hypothesis to explain vector-parasite specificity (N~uttall. 1908) which re-
mains an attractive explanation. Digestion of the blood meal by mosquitoes
begins at the outer edge of the meal and proceeds inward (Davis and Phillip,
1931). This has been observed in C. pipiens (Hfuff, 1934), A. aegypti (Stohler,
1957), Ain. st-ephkensi, An~. goinbiac and Ainophales macutlpipenns M4eigen (Frey-
vogel and Staubli, 1965). Before engorgement they become squamnous w~ith con-
vex internal borders (How~ard, 1962). Bertram and B-ird (1961) observed in
the midgut cells of Aedes after a blood meal, that w~horles of granular endo-
plasmic reticulum condensed around the nucleus unfold to form a complex
system "ramifying throughout the cytoplasm" which revert to whorles when
blood digestion is complete. These changes, confirmed by Staubli et al.
(1966), apparently facilitate the liberation of proteolytic enzymles for
digestion of the blood meal, and Hector and Freyvogel (1971) noted that whorl
formation wdas found only in female mosquitoes.
The anterior midgut cells of Ain. garmbian secrete fair-ly large amounts
of mucous-like material within 7 minutes of a blood meal. Al-though this
mucous forms a plug at both ends of the stomach and of ten completely surrounds
the meal, Freyvogal and Staubli (19635) have concluded that it exerts no
important effects onl d-igestion.
The presence of a! microfilaricide in the dlige-stive secretion of the
mosquito host has yet to be verified. Duncan (1926) de-termined thle bac-
teericidal activity of the gut contents andl feces of several bloodsucking
ar~thropods includilnD d~l-,haeled bi/'ureatusi (L~innaleus), A. cineran,~i~ StOmoge~~i
caleitr~ans (Linnaeus) and blood fed Muse2~r d~omesaican (Linneaus). Activity
was found against 8 of 18 species o-f bacteria used. The bac trial material
fromi S. calcit~ranti was heat stable (1000C for 30 minutes) and wias not
destroyed by trypFsin. St. John et al., (1930) found no evidence of bac-
tericidal material in the digestive tract.
Kartman (1952) also considered that th~e sped of microfilarial passage
thr-ough the midgut to be of critical importance to lar\'al survival, and was
probably related to the pr~esence or absence of salivary anticoagulants.
Salivary/ anticoagulants have beenl d~emonstrated in Ani. m:iaultipei:s (Yorke
and I-acfie, 1924i), An2. qluadrimaCulatusln and Aln. pun~chipennis (Mletcalf, 1945),
AnzophcZes subpicius Grassi (Cornwell and Patton, 1945). They are absent in
Ai. aegypriti (Mletcalf, 1945), Aiedec den!tL'itls Haliday (Shute, 1935), An,.
steph~ensi (Shute, 1948), C. quin~quefa~isciatus!: and C'. pip,~ien (Shu~te, 1936).
Salivary anticoagulins appear to prevent the premature clotting of the
blood me~al (Lloyd, 1928).
Salivary aggiutinins are absent in A. ccepy bi, A?. dense ~'- jo: (S~hute,
1935), As stephens,-I: (Shutle, 1935), and C. p:,ipiere (Shute, 1936). Sal-ivary
aggjlutinins have been demronlstrated in~ ::. r::zu iZir,:niic (Shlute, 1948), Ajn.
quadvimou7 abnL 1s (Metcalf, 194;5), and Anii. subpi:-1- as (Cornw~ell an i Pa tton,
1914). It should be noted that salivaryj aggllutininr -fromr An. quadrimacu-
Latu:: are effective against red cells o amlcw idg abt
and mouse, but not chicken or turtle (Metcalf, 1045).
The addition of an anticongulanlt to the infective blood mieal produces
a rate of microfilarial migration in I. na:7Y~ti, a poor vector, which is
quite comparable to that in An,. quadrZjima~nu7latus, an efficient hos~t (Kartman,
1953). Zaini et al. (1961) reported 911 of B. molayi microfilariae ingested
by Aedes F'inl~a::a) t~oloi (Theobald), which lacks an antiicoagulin, were trapped
in the stomach blood clot after 24 hour~s.
The speed of blood digestion by the miosquito may affect mnicrofilaria
survival and is often quite variable between species. Variations in digestive
rates are related to such factors as photoperiod (0'Gower, 1956), temperature
(Sella, 1920), season (Guelmino, 1951), and the amount of blood ingested
(Deeoissezon, 1930). De Buck et al. (1933) suggested that refractoriness of
Anr. macu-Zipennis to Plasmodiumi was correlated w~ith slow digestion of the
blood meal in certain varieties during overwintering.
Thayer et al. (1971) have advanced the interesting hypothesis that host
speci ficity of the insect and its abi lity to transmi t certain parasites may
be related more to components of the blood meal itself rather than differ-
ences in digestive power. In a comparison of the free amino acid levels
following a human blood meal to those after an avian blood meal revealed
significant differences in the concentrations of aspartic acid, isoleucine,
One of the most consistently demonstrated phenomena regarding thre
resistance of mosquitoes to invadiny microfilaria is that of encapsula-tion,
a process by which the parasite becomes imnpregnated w~ith a brown deposit
which eventually forms a capsule. Encapsurlation or chitinization, is
currently viewed as a defense mechanism (Salt, 1957), which may occur
around living or dead microfilaria (Kartman, 1950). Yen (1938) reported
encapsulation of all larval stages of D. immitle in the stomach, Malpighiani
tubules, and the body cavity of mosquitoes. Schacher (1962) reported the
recovery of encapsulated larval i. pahan-.iS from Anr. qarjn!imaonuZabtus and
Salt (1955) reported two defense reactions of M!icrolepidopteraer against
experimentally injected parasitaid eggs, which consist of encapsulation and
the deposition of melanin. The deposition of me~lanin, a well-known reac-tion
of the tyrosinase complex of enzymes on tyrosine, or an intermediate chromo-
gen in the presence of oxygenr, acts as defense reaction only fortuitously,
when the deposit is so cited as to prevent a vital activity, such as the
hatching or the feeding of the parasite (Salt, 1957). A similar reaction
was noted in parasitoid-infected Carausinc in which Salt (1956) described
the melanin reaction to be as strong in confined spaces as in the main blood
stream and progressing to encase the parasite in a brittle black sheath.
Salt (1960) described a rapid reaction of the haemocytes of the tomato moth
caterpillar against infected parasitoid eggs or larvae, which led to encap-
sulation within 4 hours and the subsequent death by asphyxiation of encapsu-
lated parasites. The haemnocyte reaction was similarly invoked by other
species of' parasites, by dead parasites, by glass rods, pieces of nyloni
thread, by injured pieces of the caterpillar itself, and by organs trans-
planted from other species of caterpillar. Salt (1960) concluded that sur-
face properties of the parasite or implant determine whether or not it will
be encapsulated. Goodwin (1958) previously noted that the action of diethyl-
carbamazine on microfilariae and diaminodiphenoalkane derivatives on schisto-
somies was "to change the surface of the parasite in w:uch a way that it becomes
recognizable to the host as an intruder."
Bronskill (1962) noted an iden-tical reaction in the larvae of A1.
an:!Urpti, ieAes schautb ans (Walke-r), and AndecZ twic~hu!ri. (Dyar), wh!ien chab-
dita-id DD136 juveniles (an undescribed species of the Nieoaplactanida? with
its associated bacterium) penetrated the blood sinus and entered thle body
cavity. By 5 hours, a thick capsule developed about many of the ensheathed
imm~ature forms as a result o-f a rapid defense reaction of the mosquito which
had both a melanin and a cellular manifestation. He further noted that this
type of host resistance resembles that of adult mosquitoes to filarial nema-
todes. Poinar and Leutenegger (1971) have recently shown that this reaction
in mosquitoes is a humoral type of melanization and unlike typical melani-
zation reactions in insects, does not depend on the direct presence of
hemocytes as a precursor stage of melanization. They describe a deposit
which slowly became pigmented, originated from components in the noncellular
portion of the hemolymph that coagulated out on the surface of 4 neoplectanid
nematodes in C. pipiens.
Whether the "gut barrier to infection" is due to the enzymes of the
digestive processes or to mechanical factors of movement through the ali-
mentary tract remains unknown. Chamberlain and Sudie (1961) have advanced
several hypotheses including virus inactivation by digestive Fluiids and the
impermeabili ty of the peritrophic membrane, but emphasize that no mechanism
has been completely proven. Lavoipierre (1958) reviewerd the relationships
between filarial nematodes and their arthropod vectors including a brief
discussion of the possible role of digestive physiology of the vector in
limiting the intensity o-f infection.
The endocrine physiology of thle arthropod vector has also been impli-
cated in the mediation of filarial competence. The insect endocrine system
is remarkable in its complexity and its ability to integrate development and
physiological events wi~th the environment. The pdfaSitiC nematOde mUSt re-
spond appropriately to this changing mnilieu of hormones because they reflect
the external environment in which the host finds itself. Davey and Hominick
(1973) observed that where the nematode-arthropod association is obligatory,
thle long evolutionary history of host and parasite mnay have resulted in a
dependence of the nematodes on the? endocrines of the host. They suggest
that such a dependence might take the form of host endocrines providing cues
to the nematode to sense the hiormone molecule itself; an indirect cue would
have the nematode respond to a hormone induced change in the physiology of
Numerous reviewsof insect endo~crinology have appeared recently in the
literature (Highnam and Hill, 1969; W~iggleswvorth, 1964, 1970). At least
6 enldocrine tissues have been described in insects. Neurosecretory cells
of the pars inte~rcerebralis of the brain have axons which terminate in the
corpus cardiacuim, a storage and release organ closelyv associated with the
heart. These cells secrete "activation hormone" which activate the thoracic
(prothoracic) gland (Wigglesworth, 1970), and also a factor that is involved
in the control of egg formation (Davey, 1965), protein synthesis, and the
mocbilization of lipids (Wigglesworth, 1970). The hormonal cycle in blood-
suckers is initia-ted by distention of the m~idgut wiall following engorgement
(Certram and Bird, 1961), wJhich stimuilatess the neurosecretory cells in the
braini to release a hormone into the~ hem;ioccel (Lea, 1967). Thiis in tulrn
stimulates the corpuis cardiacum to secrete the gonadotropic hormone; oocysts
begin micropinocytossi (Roth and3 Porter, 1964; Anderson andl Spielman, 1971),
vitellogenin is synthes-ized (Hage;lorn and Judson, 1972), and eggs subsequently
mature. The corpus cardiacum also contains its own intrins-ic cells which
produce and release a variety of other hormonal factors chief among them a
peptide active inr the con-trol of he!art andl other muscles (Davey, 1964).
The paired corpora allata produce juvenile hormone (Wigglesworth,
1954, 1964, 1970) which is active in development, and is essential to the
full expressioni of egg produc-tion. Juivenile hormone dietermines the dir-ection
of the development initiated by ocdysone; whether toward another larval form,
a pupal form, or the adult form (Wigglesworth, 1954, 1964, 1970). It also
governs the titre in the hemnolymph of the yolk protein by regulating its
synthesis in the fat body (Engelmnann, 1969). Juvenile hormone has been iso-
lated from a wide variety of plant and animal sources (Schneiderman and Gil-
bert, 1958, Wlilliams, M~oorhead and Pulis, 1959). Schneiderman, Gilbert and
Weins-tein (1960) reported that cell-free broth of Eacherichia~ coZi and whiole
broth from Proteus species had a very high juvenile hormone activity in the
The principle remaining endocrine organ, is that thoracic gland which
through the production of ecdysone or moulting hormone, initiates those
developmental events culminating in the production of ecdysone or moulting
hormone. It also initiates those events culminating in the production of a
new cuticle and the shedding of the old. Although insect gonadal tissue
has been considered as nonendocrine, in somne insects at least, the testis
(Naisse, 1966), and the ovary (Pratt and Davey, 1972) have been shown to
produce hormones. Even the hinrdgut has been implicated as a source of hor-
mones in at least 1 species (Beck and Alexander, 1964).
The role of hormo~nes in m~osquitoes is of special interest since
Bhattacharya and Chowduiry (1964) reported that certaini vertebrate steroids
appeared to increase the susceptibility of A. aegypti7: to infection with
w. banicrofti. Spielman and others (1971) reported that 1 steroid hormnone
ecdysteronie, stimulates ovarian development in adult mosquitoes, an effect
not observed after the administration of various vertebrate steroids. Howi-
ever, administration of massive quantities of ecdysterone did not influence
the development of B. Pahan)1i. Gwai~rdz and Spielmnan (1974) reported that
application of synthetic ecdysonea~nd juvenile hormone to AI. aegyabti did not
affect the development of Bi. pahiang-i.
Certain oxyurid parasites of roaches appear to require hormones from
the hosts' corpora allata (Nadakal and Nayer, 1968), and median neurosecre-
tory cells (Gordon, 1968, 1970) for developmennt. Even nonentomophilic nemia-
todes may be influenced by the presence of such exogenous insect hormones
as juvenile hormone and ecdysterone (Hitcho and Thorson, 1971; Johnson and
Viglierch~io, 1970; M~eerovitch, 1965; Shanta and Me~erovitch, 1970; Wiebster
and Craig, 1969; Davey, 1971).
Davey and Hominick (1973) have concluded that however attractive the
notion that development of a parasitic nematode is under the control of the
host endocrine system may be for many authors, the cases in which this has
been demonstrated are very rare. Yoeli and others (1962) found that micro-
filariae of D. ilnv~itiS developed to the sausage stage in Al. qualtimac=ulatus
decapitated immediately after the infective blood meal. Since the vitello-
gen-ic hormone is secreted and stored in the head of the mosquiito, this find-
ing established that the inritial stages of larval differentiation are not
con-trolled by the same hormone systemn that regulates host oogenesis. Gw~ardz
and Sp~ielman (1974) showedc thati the allatu:n secretion does; not influence
the development of larval B. pshangi. Such larvae3 nature in allactomized
juvenile-hormone "free" females of both suisceptible and refractory mosquito
species. These authors also concluded that larval development of B. paFhangir
wa~s independent of the insect hosts' enldocrine milieu.
Dual Etiology of Parasitic Diseases
The relationship of normal intestinal microbiota to the vectoring
capacity of insects is an idea that has not yet been adequately investigated.
Insects are not particularly unique among animals in their relationships in-
volving bacteria and fungi. Parasites passing through the host's alimentary
tract are exposed to both the microorganisms that reside there and to the
intestinal climate which may, in part, be influenced by them. Indigenous
microorganisms present at the host-parasite interface must be regarded with
as much consequence as the host itself.
Gnotobiotic studies have shown that the pathology of many parasitic
diseases depend not only upon the etiological agent, but also upon the host's
intestinal microflora. Phillips and Gorstein (1966) have repor-ted that the
association of ameba Fntamoeba histotytica Schaudinn with living bacteria
such as Escherichia coZi (Migula), Aerobacter aerogenes (Kruse), and
Clostridiumn perfringens (Vellon and Zuber) was necessary to produce typical
hepatic abscesses in hamsters. Wittner and Rosenbaum (1970) confirmed this
study and observed that the participation of bacteria with ameba must be
more than just maintenance of a suitable intestinal environment, but might
involve a resistance transferring factor described by Anderson (1965) passed
from bacteria to ameba. Phillips (1964) considered that the bacterial flora
acted by "providing a suitable environment, physical and chemical, for excys-
tation and establishment of lumen infection until such time as the ameba can
enter the tissue."
Infectious enterophepatitis in turkeys has been shown to reflect a dual
etiology requiring Histonomas moeoegridlis (Smith) and a single species of
bacteria (E. co~i, C. per~fr~ingeno, OF Bacilluts subtilis Cohn) (Bradley et al.,
1964, Bradley and Reid, 1966). Franker (1965) produced a moderate rate of
infection with this protozoan and a species of infection w~ith this protozoani
and a species of Eimar~in in comb~inationn with 3,~ZwIn C~u a ciru Frankland and
FrankZland or I~actoba~cithe fe medir;~zi. Deijerinick. Springer et al. (1970)
noted that the bacterial requirements for producing this disease in bacteria-
free chickens were different from those for the disease in bacteria-free
turkeys. The essential contributing factor of bacteria in the pathogenesis
of infectious enterohepatitis was determined to be neither a favorable pH,
nor an oxidation-reduction potential value within the intestine of the host,
but rather th make the cecal environment more suitable for survival of the
transport host, liateratkis gaZZinavuzm (Schrank).
Wlescott (1971) noted in gnotobiotic meice infected with H~ematospir2oide~s
c-1ubiuso Baylis and Nirppostrongyi~ue brasilctie is ravassos that mo~re parasite
developed, infections were of greater duration, and of a higher reproductive
potential in the conventional rather than in germ-free hosts.
Feline infectious enteritis was shown by Johnson and others (1967) and
Rohovsky and Griesemer (1967) to be almost asymptomatic in germ-free cats,
but severely symptomatic in speci~fic-pathogen-free (SPF) cats. Visco and
Burns (1972) reported a parallel situation in gnotobiotic chickens infected
writh Eimenia tenclza (Railliet and Lucet). It w~as previously demonstrated
by Radhakrishnan (1971) that E. tenetta could produce typical cecal pathology
in qnotobiotic chicks only wihen associated with certain bacteria or combi-
nations of bacteria. Johnson (1971) noted severely depressed growth and
development of Alccurid2ia gazZEi (Schrank) in germ-free chicks and Bosmak
(197'1) reported that Hy't"na~no!;lpic: nanae (von Sifbold) in mice appears depen-
dent upon host bac~teria, and that the variety of bacteria present changes
withl increasing intensities of tapeworm burden. Conversely, the absence
of bacterial associations may be a factor in disease transmission. For
successful transmission of leishmania, the PhZeboornu.; fly must have a
sterile alimentar-y tract because human leishmanias do not tolerate bac--
terial contamination. Infection of A. stophensi mosquitoes with a micro-
sporidian parasite Nosemae al~gorae Vavra and Undeen, 1970, reduces their
susceptibility to infection with the simian malarial parasite, Plasmodiumn
cynomolgi Mayer as measured by mialarial oocyst counts 6 days after the
infective meal (Ward and Savage, 1972). Mortality from nosematosis was so
great that 90-95% of the exposed mosquitoes died by sporozoite levels
could be assessed. Fox and Weiser (1959) pointed out that there is
apparently no evidence that Njosema is physiologically antagonisitc to plas-
modium, or that the former attacks the latter. These investigations con-
cluded that the miidqut wall of heavily infected mosquitoes was so dis-
integrated that suitable sites were not available for oocyst development.
From these and similar studies, it is apparent that the intimacy of host
bacterial flora with parasites appears to be more significant than cir-
Microflora Associated with Mosquitoes
Mitchell's (1907) report that larval A. aegypti were predominantly
bacteria feeders was based on the observation that larvae develop more
rapidly in water contaminated by sewage. The same year Putter (1907) re-
ported that organic colloids and solutes serve as the primary food source for
aquatic invertebrates. Hinmlan (1930, 1932) andi Trager (1936) considered
that var-ious substances in solution ser-ved as the sole food sour-ce, buit
Shipitzinia (1930), and Rozebloom (1935) and Buddington (1941) showed
that pond water passed through a bacterial filter- could not suppor-t growth
beyond the further instar. Trager (1935), however, reported that larva could
utilize liver extract and whole killed yeast in solution.
Lamborn (1921) claimed that some mosquito larvae feed on specific
living organisms suchl as Si~irro:7r~a sp., Eugnl~an sp. and JVolvox~ sp., but
M~etz (1919) showed that it made little difference whether a variety of animal
or vegetable matter fed to anopheline larvae was living or dead. Hi nman
(1930) reported that bacteria, algae, and protozoa although ingested by
mosquito larvae, were not digested and hence, contributed little to the
nutrition of larvae.
Laird (1956) concluded that algae and protozoa are the primary sources
of food for mosquitoes breeding in permanent ponds and lakes in mosquito
breeding areas of the South Pacific. Coggeshall (1926), Hamlyn-Harris (1928),
Senior-White (1926), and Howland (1930) have all reported that algae is the
primary food source of at least some mosquito species.
Atkin and Bacot (1917) raised larvae of Al. aegypti in pure cultures of-
Sacchar~omyces cerviiciae? Meyen and Ba-iZllus coZi courauniior (Dur-hamn) and Bacot
(1917) noted that mosquito larvae kept in water originally turbid with bac-
teria, was rapidly cleared. The intestinal contents of such larvae were
relatively free of bacteria, which Bacot (1917) considered a result of com-
plete digestion. GldSser (1924) considered that liver and yeast supplied
certain accessory growth factors to the larvae of the several species of
flies he maintained in culture. Bragina (1926) nourished newly hatched A.
menuipenison bacteria (unidentifieId) and Barber (1927) found that algae
alone, bacteria alone, or infusoria alone consti tutes a suffricient food
source for Ainopheles? cru ;ianci andi A. criuadei:;:culatus.n Thlese infusorinsi~
wer-e chosen as r-epresent;ative of the or~ganisms commonly found in the plankton
of Anloph~els-producing water and included: Cotpidiumn sp. isolated from
cultures of rotting algze, a motile, unlicellular, grass-green algae, possibly
ChiZ~mndomnonacs sp. or a related form:, anld a lar-ge and a small variety of~
SpirlZElum sp. Barber (1928) reported similar results using these organisms
as food for culicine larvae.
Rozeboomn (1935) reported that bacteria, to a certain extent, can be
utilized as food by A. aegypti, C7. pipien~s, Ciculex te2riritnr s Walker and
Cu~e- salinaruss Coquillet larvae. Environmental bacteria, associated writh
the different natural breeding places of the mosquitoes, promoted optimum
development when bacteria were the only source of food. Rozeboom (1935)
found that E. coli, Bacilluis subtilis Cohn, BaciZllus myrcoides Fluigge, A.
aetrog~enes, and Pseud~omonays aeiruginzosa Migula were of equal value, but that
Sarciina Zubea Schroeter was inferior. Pseudomnonas fluoremen3 s M~igula had
a toxic effect on young larvae. Bridges (1965) reported that B. subtiis,
Bi. mega~therium, and Staphylococcus aureuls Rosenbach were unsatisfactory as
a sole source of food for C'. qu~77inquef~j~atlscau larvae. The yeast S. ce?e-
visiae provided the best food source. E. colL and A. aerogenes w~ere about
equally indifferent in food value. P. aeruginosa and S. Lusta were utilized
as food only to a limited extent. Buddington (1941) reported that E. co~i,
B. su~btilis, and BaCiZtuS meantheriumr DeBary in B~erk~efeld-W filtered pond
water supplied sufficient nutrients for the complete development of Ai. ac,~-
Beattie and Howland (1929) considered that bacteria play some part in the
nutrition of orthopodomyia sp. and Rudolfs and Lackey (1920) reported that
microorganisms fr-om decomposed vegetable matter w~ere responsible for miosqua;
breeding. Laird (1956) concluded that bacteria were the primary food of
rain pool mosquitoes because bacteria were thle most abundant organisms !1row1
ing in these transient ponds.
Yeasts have been shown to contain most of the nutrients needed by
mosquitoes for complete growth (Atkin and Bacot, 1918; Trager, 1935; Fred
et al., 1936; Subbarow and Trager, 1940; Buddington, 1941, Goldberg and
de Meillon, 1948; Akov, 1962). Alcohol sterilized, but not heat sterilized
yeast will support complete development of mosquito larvae (Buddington, 1941).
An artificial media supplemented with autoclaved yeast has been used to rear
normal adult mosquitoes (Lea et al., 1956; Akov, 1962).
Infusions of organic material have been used to successfully produce
adult mosquitoes of many species (Atkin and Bacot, 1917; Barber, 1927;
Hertig, 1936) in the ovaries and testis of C. pipiens. Brumpt (1938) described
Ricketecia curlicis Brumpt which had been fed 12 days previously on a subject
with filariasis. Sellards and Sites (1928) frequently noted masses of ricket-
tsiae in the lumen and epithelial calls of the hind gut of Ai. aegypti.
These rickettsiae were present in specimens known to be infected with dengue
fever virus, whbereas they wvere not observed in control mosquitoes. More
recently, Venters et al., (1971) reported the presence of rickettsia-like
organisms in the midgut cells of 25% of A~n. stephe~s~i examined. Since they
observed such organisms only in stomachs free of malarial cocysts, they
considered mosquito susceptibility to malarial infection can be influenced
by the presence or absence of this organism. Yen and Barr (1971a) considered
that the cause of cytoplasmic incompatibility in C. pipiens was due to
rioZbachia-like rickettsia which are strain specific and localized in the
follicles of the eggs. These microorganisms, especially abundant near the
mnicropyle of recently laid eggs, lie in close proximity to sperm cells entering
the micropyle in their passage to the femlalee nucleus (Yen an~d Barr, 1971b).
These workers suggest a deleterious effect by these microoi-ganisms on pre-
sumably incompatiblel" sperms so that they are unable to participate with
the oocyte nucleus in -the -formation of a zygote. Rickettsia-l-ike organisms
are not novel to blood-sucking arthropods, however, and new x~amlpless continue
to be reported in such insects as Rhipicephp~alus bursa Canestrini and Fan-
Zdngo (Friedhoff, 1970).
The bacteria associated writh mosquitoes have only begun to be docu-
mented. Star and Micks (1957) (quoted by Ferguson and M~icks, 1961) isolated
22 different bacteria cultures from 3 to 5 intact adult females of c. fa~tigans~
and 35 different strains from a homogenate of 49 specimens. Eighty percent
of these were Gram negative rods.
Hinman (1932a)found bacteria as w~ell as yeast within the eggs of
A. aegypti and Kellen and W~ills (1962) reported evidence of transovarial
transmission of Thetlobanis sp. in various species of California mosquitoes.
Chao et al. (1963), however, reported their failure to isolate microorganisms
from within the eggs of 2 species of Cu!Ze, 2 of Anopheles and 2 of Aedes.
Mlany workers have since reported the successful axenic rearing of adult
mosquitoes from surface-sterilized eggs (Dougherty, 1959; Akov, 1964; Nlayer,
1966; W~allis and Lite, 1970, Rosales-Ronquillo et al., 1972, 1973).
Chao and Wistreich (1959) isolated bacteria from the midguts of 46
adult CuZex barsatis Coquillett selected from a colony fed on apple slices
for at least 2 days after emergence. Results of these isolations showed the
presence of Allcatigen~es gutactus (Zimmerman). Achromobarcter ap., and Aero~-
bacter clocae Jordon, Eschzerichia Intermedia W'orkman and Gillen, F'lavn-
bacteriumn sp., M~icrococcus sp., MsocrococcuN vivi~ano Nligula, Pr~oteus sp.,
and Pseud~omonaa sp. Cultures of unidentified Gram-negative rods as well as
Sacchnromlices sp. and Geotirichum sp. were also reported. Chao and Wistreich
(1959) also noted a higher ratio of midlut sterility in males than in females,
and reported that males, after a forced blood meal could not digest the blood
and died shortly after feeding. Wistreich and Chao (1960) isolated bacteria
from midgut sections of 14 fourth-instar larvae of c. tar~salis that had been
raised in fresh tapwater and fed brewers yeast. Results of these isolations
showed the presence of Ach,?;~roobact:er sp., L~a-tobaci~Z~ue sp., Mi:cro-coccu
candidus Cohtn, Micrci'7OeCuS sp., Pr~oteus: rattge~i Hadley et l., and
P7seudomonas sp. Again, cultures of unidentified Gram negative rods and
Sacchar~omyces sp. and Geotrichycrm sp. were observed. Isolations from the
larval rearing water showed the presence of A. eloacae, Klebcietta sp.,
Sacchacromy:~cetee sp. and Geotrichim sp. Fourth instar larvae were found
to contain examples of all the genera of bacteria isolated from adult C.
tarsatiis w~ith the exception of Aerobraectr sp., Escherichia sp., and Flanvo-
bacte~riu~m sp., whereas isolations from adults did not show the presence of
Lactobaciltlus sp. or the coral pigmoented form of Micr~ococcus sp.
Bacterial isolates from the midguts of adult C. quinquefasciatus fed
apple slices for 2 days after emergence showed the presence of lch~romobacter
sp., Flavobacteriium sp., KlebuieZla sp., Mi~croocus ocaseoi~ytious Evans, and
Pseudomonas sp. (Chao and Wistreich, 1960). Midgut sections from 11
fourth instar larvae of the same species produced only BacitZus sp. and
the yeast Saccharonyces sp. No apparent difference in the microbial flora
from male and female adults were noted, with the exception of Miicrococcus
sp. and Saccharomyices sp. which were isolated from female mosquitoes only.
Wistreich and Chao (1961) extended their research to microbial iso-
lation from larvae and adult Aetdeos sierrensis Ludlow and Anzopheles albimialus
W~eidemramfed on a larval diet of yeast and adult diet of apple slices. Lar-
val isolation fr-om A. a~hbimanue Weidemarnfed on a larval diet of yeast and
adult diet of apple slices. Larval isolation from A. athtman~us showed the
presence of Alcaligecnas sp., Racsittue and unidentified Gram-negative rods.
Adult A. aZbim~anusi showed Ach:~i~r,~romebactrsq>., Ahealigena s sp., and corynle-
bacterium sp. Saccaharamyces sp. w~as isolated from both larvae and adults.
A. siervenclsis larvae contained Bu~illus circuzlanoi: Jordan, Paracolobact!rum?
-in-tennedi~um Barman and Prettger. Alchelomobac~ter sp. and E'. inte~~irmdia were
found in adults. Sach:aromyjces sp. were isolated From the larvae. W~istreich
and Chao (1961) also reported that microorganisms isolated from the adult
and larval environment were distinctly different from isolates of the mid-
Wistreich and Chao (1963) reported microbial isolations from larvae
showed the presence ofl the following bacteria: AZealigenes sp., Coryne-
barcterium sp., Escherichia fraunditii (Braak<), Ku~irthia beasoni Hauduroy et al.,
Sarcin~a flava de Bary, as well as fungi belonging to the general Geotrichzum,
Penzic~iumi, anld Saiccharomylces. Isolations fromn adult mosqjuitoes showed
the presence of A~chromobacter sp., Alcaligenes sp., and Saicca~romyces sp.
Furthermore, they reported an especially high rate of midgut sterility in
the adults of this species. Microorganisms were isolated from 5 to 21 fe-
males and 4 of 9 males. Alealigenes sp. wiere found only in adult females.
Ferguson and Micks (1961a) instituted a similar series of experiments
using adult female C. fatigans from a colony that had been denied access to
food. By means of a sterile riicro-dissection technique, they demonstrated
the presence of LactobacEiZtus sp., AlZealigneszc sp., PoeudZ~omonas sp., and 2
unidentified cultures of Gram-negative rods. Bacterial isolates from thle
midgut o-f similarly tr-eated adult female Ani. quandr;;maculatuszi proved to be
2 strains of Strenoto-ooccls sp. and a species of Ac~~orobcter (M'icks and Fergu-
son, 1963). Ai. accypt~i': produced only a species of CorynobaSrcctr~iwnj and no
microorganisms could be cultured from the mnidguts of adult female CutZe:c
The midguts of newly emnergjed adults of i'. frtlatigan, C. m~les~tus, An?.
quadrj,,iimacultus. and A. aegypliti which w~ere not allowed access to either
blood or sugar were examined by electron microscopy (Micks et al., 1961a).
Ultrathin sections of C. fa~tigans midgut revealed the presence of intra-
cellular microorganisms which M~icks et al. (1961a) described as Gramn-
negative rickettsiae. Since these organisms did not appear to produce
pathologic changes within the cells and were occasionally found in mycetome-
like bodies. Micks et al. (1961a) suggested they might be symbiotes.
The question of a symbiotic relationship between mosquitoes and
microorganisms has yet to be satisfactorily answered. Brooks (1965)
generalized that arthropods possess symbiotic microorganisms only if they
feed on nutritionally inadequate (i.e., incomplete) diets during their en-
tire life cycle. Blood or serum is considered inadequate since it is
deficient in B vitamins. Furthermore, since mosquitoes undergo complete
metamorphosis from a larval stage spent as an omnivorous or scavenger-
feeder, they are considered to pick up an adequate amount o-f vitamins and
intestinal microflora (Brooks, 1964) in that stage to suffice for their
entire life. Blood-feeders with incomplete metamorphosis which feed on
blood during all immature stages, do have symbiotic microorganisms present,
and as a rule, their alimentary tracts are otherwise sterile. Such arthro-
pods are represented by bed bugs, sucking and chewing lice, and ticks and
mites. Removal of symbiotes by axenic rearing procedures causes severe
growth impairment, body malformations, lack of reproduc-tion and death
(Baines, 1956). Wdigglesworth (1929) considered that symbionts provide some
accessory food sutstances to tsetse flies that are similar to the vitamins
necessary for mammalian growth. B vitamin supplementation of the blood
med1 fed to axenic blood-feeders does, in fact, alleviate the growth factor
deficiency but fails to restore normal reproduction (Pucha, 1955). Aschner
(1934) found that sumbionts are a source of food for the body louse Pediculus
hom~inus hominu:- Linnlaeus. Young lice died within 5 or 6 days when deprived
of the mycetomes that contain symbionts. Actinemy~cr e I'hodnri~ Erickson in
the gut of Rhodninea sp. are essential for the maturation of the insect
(Brechner and Wigglesworth, 1944).
In order to obtain further evidence of a possible symbiotic relation-
ship between C. fabtigans and its associated microorganisms, M~icks and Fer-
guson (1961b) placed a number of females on a chloroamphen-ical -dihydrostrep-
tomy~cin-sugar solution diet -for periods of 5 to 12 days. Mosquitoes relieved
of their microorganisms by such means were subsequently allowed to take a
blood meal from canaries infected writh Plasmodium~ relictum Grassi and
Feletti to ascertain possible effects on host immunity. Antibiotic treated
females became inac-tive, lacked coordination, were unable to fly, and when
dissected, their midguts were found to contain large masses of undigested
blood as long as 10 days after the blood mieal. The normal rate of digestion
of human blood, although variable with -the species and day--length, is re-
ported to be 31 to 48 hours (0'Gower, 1956). Eighty-five percent of the
treated mosquities and 74% of the untreated controls were infected with P.
2elictuw. In general, antibiotic treated individuals contained about twice
the number of oocysts per midgut as did controls.
M~icks and Ferguson (1961b) concluded that microorganisms play a
significant role in the susceptibility or resistance of the mcosquito to the
mnalarial parasite. They mention the possibility of competition between the
miicroorganisms anid the malarial parasites for essential nutrients, in w~hichi
case, the antibiotic would work in favor of thle P-lasmod~iumi. Alternatively,
they suggested that certain strains of these m~icroorganisms could supply
factors needed by the host for metabolic processes and defense mechanisms,
and that killing thle organisms increases host susceptibility to malaria.
Unfortunately, their work appears to have been discontinued at this stage.
Furthermore, the feeding of antibiotic diets, besides killing mrost, but not
all, internal microorganisms, may have produced additional e-ffects on host
physiology not considered by these workers.
Shymala et al. (1960) reported that chloromycetin in the diet of the
silkworm influenced the digestion and utilization of protein, fat, and
minerals. Tetracycline, oxytetracycline, and chlorotetracycline have been
shown to combine specifically within the mitochondria of living monkey kidney
cells and with the cells of salmon?1c:~ byphoc-a Wlhite (Dubuy and Showracre,
1961). Although the mammalian cells remained alive for days, there is much
that remains unexplained about the effects on metabolism and nutritional
requirements of antibiotic-treated cells.
The roles of microbiotic symnbiotes within mosquito midguts are manifold.
Hinson (1933) found within the bacterial cell, a factor which stimulated
growth of mosquito larvae. Rozebloom (1935) considered that different
species of bacteria possessed varying nutritional properties for mosquitoes.
F. coli, B. subtii's, B. mi~eoides and P. flo2scens were equally satis-factory
as a sole source, while S. Zutea and Pseudomonas pyeaneta Migula (P. aerug~inosa
Schroeter) were detrimental. Arnal (1950) considered that specific symbiotes
were essential for the digestion of erythrocytes in mosquitoes and described
intracellular organisms in midguts of C. p~ipien which secreted enzymes
capable of hemolyzing erythrocytes. Micks and Ferguson (1961b) found un-
digested blood only in those mosquitoes in which the microbial flora had been
reduced or eliminated by the antibiotic diet. Ferguson and M~icks (1961b)
also noted that one of the bacterial str-ains isolated from the midgut of
C. fatigans produced hemolysis of human blood. The work of Chao and W~istreich
(1959) support this relationship between midgut microorganisms and the proper
digestion of blood. They found a higher ratio of midgut sterility in males
of C. tarcatio and also noted that males could not digest blood with which
they had been fed. Terzian et al. (1952) reproduced the "indigestion"
syndrome in A2. aegypti: by feeding calcium and calcium-antibiotic combinations
before and after engorging. Only diges-tion of the hemoglobin fraction of the
blood was inhibited. The effect of the antibiotic (they reported) was to
potentiate the inhibi-tory effect of the cations upon the particular enzyme
system involved. Terzian (1953) also reported that streptomycin and chloro-
mycetin increased susceptibility of Al. aeigypti to Plasmodium gallinacewr!
(Brumpt). The author apparently did not consider that this increase in sus-
captibility might be due to antibiotic mediated alteration in host microbial
Statement of the Problem
This dissertation is a series of 4 investigations on the relationship
between filarial vector and its resident microflora in the midgut. The
participation of these microflora in the transmission of parasitic disease
is of special interest. The a priori premise is that mosquitoes acquire
their life-long complement of midgut bacteria as indiscriminate filter-
feeding larvae. During growth, competition among microorganisms establishes
a microecosystem of bacterial types which are retained throughout life.
The arthropod midgut is thle only part of the alimentary tract which is not
shed during molting. This bacterial complement is individually distinct be-
cause of its fortuitious acquisition, yet species specific because of the
consistencies of egg-laying site selection among mosquito species.
The influence of this internal microhabitat upon potential parasites
passing through the alimeintary tract may be manifold. It may be of
sufficient importance ot determine the fate of the pathogen, and thus, the
efficiency of the vector. The recently developed techniques in gnotobiology
presented the opportunity of selectively separating the arthropod vector
from its indigenious microflora.
The first problem was to titrate levels of filaria infection in the
mosquito, determine their effect on the mosquito, and the ultimate fate of
the larvae. Of parallel interest wJas the phenomenon of early death in in-
fected mosquitoes. This investigation is reported in Results, I.
Secondly, a system had to be devised whereby a very efficient filaria
vector such as An?. quadirim~aculatusl could be raised, maintained, and infected
under gnatobiotic conditions. Previously reported techniques were not
appropriate to this species. Results, Section II describes this effort.
The problems of defining the "normal" bacteriologic flora of mosquitoes
was also unique. Nlo reports of bacteriological evaluations of wild-caught
mosquitoes have appeared in the literature. Furthermore, such reports as
have appeared, were not quantitative. Results III reports a quantitative
and qualitative evaluation of 5 groups of wild and colonial mosquitoes.
Finally, the effect of the normal flora on the establishment of
filarial larvae in a susceptible mosquito is investigated. Pure cultures
of bacteria isolated from wild mosquitoes are "added back" into gnatobiotic
mosquitoes as monocontaminants. These mosquitoes, as well as conventional
and bacte ia-free specimens were infected w~ith D. immitis larvae.
For-ty-eight hours later, all mosquitoes were terminated and the degree of
filarial development assessed. This is reported in Results,Section IV.
I. BIOASSAY OF EARLY VECTOR MORTALITY
FOLLOWIING DoIROFILARIA ~IMIIs LEIDY INFECTION
The m~ost sal-ient problems attending experimental attempts to transmit
dirofilariasis through laboratory mosquitoes is the initial high vector
mortality. Various authors including Bradley (1953a), Kershaw, Lavoipierre
and Chalmers (1953), Webber and Hawkins (1953), and Weiner and Bradley (1970),
have reported mosquito mortality appr-oaching 100%: in the first days follow-
ing the blood meal. A common antecedent to death appears to be distension
of the abdomen persisting 2 to 3 days, which W~einer and Bradley (1970)
interpret as a result of improper digestion of blood.
Mortality subsequent to the ingestion of large numbers of infective
larvae in the blood meal has been shown to be a result of mechanical de-
struction in the insect tissue (Pistey, 1959; Villavaso and Steelmian, 1970).
Larval destruction of the Malpighiann tubules as they leave these organs
on the fifteenth day is an obvious cause of mortality. Esslinger (1962)
documented three pathological stages consequent to infection of Anop7heles
quadvrimaouiatus Say with Brug-ia pah~angi Buckley. These include perforation
of the peritrophic membrane, destruction of the gut epithelial cells and
puncture of the basement membrane, and finally, destruction of the fat
body. However, death due to mechanical destruction after infection of low
or moderate numbers of microfloriae seems inappropriate in such successful
filarial vectors as mosquitoes are known to be.
The incidence of mosquito mortality during the first 3 days after
infection with D. inimitic may be the result of (A) mechanical action
of the larvae, (b) a toxic factor elaborated by the larvae, or in response
to the larvae, or (c) disturbances in endocrine function occasioned by
the larvae. Even subtle physiologic insults produced by filaria larvae
miay be greatly magnified in the female mosquito as all the endocrine
machinery of egg-laying begins to function.
The following bioassay is an effort to determine which of these 3
possibilities, if any, is the cause of vector mortality in the first 3
days following infection with D. immoitis larvae.
Materials and Methods
Anl. quradrimaculatus (Gainesville strain) were reared and maintained
in an insectary at the Insects Affecting Mlan Research Laboratory, Agri-
cultural Research Service, U. S. Department of Agriculture, Gainesville.
Ambient conditions were maintained at 26oC, 80% RH and 12 he~ LD. Adult
females were fed 5 days on a 2.5% sugar solution after emerging; starved
for 2 days (Terzian et al.,1957) and fed on various preparations through
either a capillary tube or a heated gutta percha membrane. Such preparations
included live microfilariae, dead heat-killed microfilariae, dead homoge-
nized microfilariae, and either the homogenized midguts, heads, or re-
mainders of the bodies of previously infected female mosquitoes.
As R.oskpi LoEs completed -feeding, they w~ere transferred to holding
cages in groups of 20 where m~ortality observations were made twice daily.
Nioribund mosquitoes were dissected and examined for the presence and con-
dition of D. immith larvae. ilicrofilariae weire collected from filuremic
dogs using the K'lein and Bradley (1973) modification of the K~nott technique.
The mnicrofilariae w~ere separated from -the cellular components concentrated
by centri-f-ugation, adjusted to a 2,000 microfilariae per 0.5 ml concentration
with normal saline (Wong, 1964). The cellular components were then reintro-
duced to bring the volume up to 1 ml. Hiomogenized material wlas sonicated
in cold saline, brought up to 0.5 ml volume with saline, and up to 1 ml
volume with canine red cells obtained from an uninfected donor dog.
The mortality of mosquitoes taking blood meals with differential
filaremias is shown in Table I-1. A concentration of 1,000 microfilariae
per ml of the blood meal permits optimum mosquito survival and results in
an average of 12 microfilariae per mosquito. Gross dissection of mosquitoes
succumbing from this level of infection revealed no notable tissue destruc-
tion in the 3 day critical period. The Malpighian tubules contained active
microfilariae in the cells at the distal end. The tubules were intact, and
entirely normal in appearance. Digestion of infective blood meals wras con-
sistently retarded in comparison to digestion in control mosquitoes.
Table I-2 shows that no significant mortality was produced b~y feed-
ing either dead intact microfilariae or sonicated microfilariae bodies to
newly emerged female mosquitoes. This finding eliminates consideration of
a toxic factor produced by, or within the bodies of microfilariae.
Table I-3 shows the effect of feeding the homogenized body parts
from mosquitoes fed 24 hours previously on infective blood meals. Neither
midgut preparations, head preparations, nor remaining body parts contain
factors which duplicate the high vector mortality experienced w~hen viable
microfilariae are fed.
Table I-1 Effects of Differential Filaremia on
Mortal ity of AnopheZen quacdiimaculatus SAY
Cumul ati ve Mlortal ity
in 3 days
Dead Intact Larva Number Females Cumulative Mortality in
(1,000 Larva~e:al) Feeding 3 Days (%)
Trial 1 24 0
2 23 2
3 20 0
4 20 0
Trial 1 30 1
2 30 2
Table I-2 Effects of Feeding Dead Intact, and
Homogenized Microfilaria to Female
Alnopheales qu~adrimaculatu~s SAY
Table I-3 Effects. of Feeding Homogenized Body Parts
From Females Fed Infective Meals 24 Hours
Adjunct to Number Femal es Cumulative Nlortal ity in
Blood Meal Feeding 3 Days (%)
Homogenized Midgut 58 5.1
of Body 73 4.1
Homnogenized Heads 20 0
Results of this study show that living microfilariae produce at
least 25% mortality in infected AIn. quadErimnr iaiulatu during the critical
first 3 days post infection. This mortality could be the result of either
occult destruction of mosquito tissue, or some factor produced by, or in re-
sponse to the microfilariae. Dead, intact larvae do not stimulate a comparable
mortality reinforcing the explanation of mechanical destruction. When sus-
pensions of dead homogenized microfilariae were fed, a comparable level of
mortality was not achieved, indicating that a lethal factor could still be
intrinsic t the mosquito. Feeding homogenized midguts of previously in-
fected mosquitoes failed to reproduce mortality suggesting the female endo-
crine system as a possible source of the factor.
Blood feeding in non-autogenous female mosqui toes initiates a neuro-
endocrine arc involving endocrines from the brain which stimulate the
ovaries to proceed with egg formation (Larsen, 1958). Neither head prepa-
rations, nor whole-body homogenates from previously fed mosquitoes contained
factors which reproduced the high mortality in newly emerged mosquitoes. An
endocrine related factor is not supported b~y these findings.
Developmental studies by Kartman (19535)and Pistey (1957) have shown
that prelarvae migrate to the Malpighian tubules very soon after ingestion,
and Pistey (1959) infers that the sudden invation of the tubules is selec-
tivity fatal. Pistey (1959) reports that the most critical period in
Div~ofitaria tenuis Chandler infected mosquitoes was the initial 3 days when
microfilariae enter the M4alpighian tubules. Intermill1 (1973), however,
describes the migration of D. irmnitis larvae to the M~alpighian tubules of
Aedec-s triseriatus Say and notes that it proceeds without any apparent
histological damage to the gut or tubules.
Inasmuch as no obvious destruction of the M~alpighian tubules wras
apparent in the present study by gross dissection, anid no toxic factors
could be demonstrated, the cause of vector mortality remlainss unexplained.
Viable, pre-larvae of D. iimmitis, however, are extremely active and subtle
alterations in Malpighian tubule function could be produced by their pres-
ence with little physical evidence. The destruction of tubule cell cyto-
plasm and cellular membranes by invading larvae has been known since the
earliest authors (Noe, 1901). Sub-tle destruction of the cellular components
of the tubules may impair nitrogenous excretion, or the retention or re-
absorption of dietary components necessary for life. Even a minute perfo-
ration of the tubule integrity could allow leakage of uric acid and other
nitrogenous constituents into the haemocoel with terminal results for the
It would seem appropriate to implicate "indigestion" as a factor in
the early death of filarial vectors. Lewis (1952) noted that the presence
of microfilaria of anchooera volou~lus Leuchlart interfered w~ith the formation
of the peritrophic membrane of Simuliumn damnosumn Theobald and Wianson (1950)
reported that blood contained in the stomach heavily infected Similiumn
which died a few days after an infective meal appeared to be undigested.
Similar observations on other filaria host combinations were reported by
Roubaud et al. (1936) and M~acKerras (1953). Lavoipierre (1958) concluded
that the most likely cause of vector death was extensive damage to the peri-
trophic membrane rendering digestion almost impossible. The peritrophic mem-
brane was impossible to demonstrate in dying, heavily infected flies with un-
digested blood meals, whereas thle membrane w~as present in controls in which
the blood meal was in an advanced state of digestion.
II. AN INTEGRATED SYSTEM FOR THE PRODUCTION
OF GNOTOBIO0TIC Azoph2eles quadriii~macuulatus SAY
A host of investigators beginning with Schottelius in 1899 have
demonstrated the fallacy of Pasteur's assumption that the host-microflora
relationship is obligate. Since then, the germ-free animal has given the
biologist one of his most valuable models to study the interactions of host
The development ofgniotabioticar~thropods was initia-ted by Trager
(1935a, b, 1936, 1937) who reported the growth of larval Aedes aegypti (L)
in sterile media containing essential nutritional compounds. In the 4
decades subsequent to that, the dietary requirements of A. aegy/pti in
axenic culture have been thoroughly studied by a number of workers (Goldberg
and DeMeillon, 1948a, b; Lea et al., 1956; Lea and D~eLong, 1958; Singh and
Brown, 1957; Akov, 1962; Lang et al., 1972). However, only a few species
other than A. aegyPti have been successfully cultured in sterile media.
They include: Cu.Eiseta -Incidens (Thom) (Frost et al., 1936), cutben pipienzs
pipicen (L) (Buddington, 1941; Dadd et al., 1973), Cutlex molectus Forskal
(Leichenstein, 1948), Aedeaz taeniorhky~chucs (W~iedeman) (Niayar, 1966), Cu1Zexn
saZin~arius Coquillett (Wallis and Lite, 1970), and Anioph~lses ctephzsna
Liston (Rosales-Ronquillo et al., 1973).
The objective of most of these studies has been to define the nutri-
tional requirements of mosquitoes, rather than to investigate the microflora-
host relationship. Of equal interest is the role of bacteria in the biology
of the host and in the pathogenesis of certain disease mechanisms.
The public health significance of the mosquito makes it one of the
most appropriate arthropods to receive such attention. An~ophecles quad2i-
malcuclatusi Say is a mlajor vector of canine heartw~orm disease in thle eastern
United States. Thils paper reports the first successful rearing of An.
quadr~imacutlatus under bacteria-free conditions and the successful use of
a lucite capsule specifically designed for gnotobiotic studies w~ith blood-
Materials and Methods
Ain. qu~adrimalcu!?atu~s eggs were collected from colonies maintained b~y
standard procedures (Anon., 1973) at the USDA Agricultural Research Service,
Insects Affecting M~an Laboratory in Gainesville, Florida. Egg sterilization
wras accomplished in 3 cm long cylinders made from the tr~uncated upper half
of a disposable plastic syringe barrel. Organdy netting was cemented over
the lower end to form a deep non-w~ettable dipping cylinder for egg handling.
These cylinders fit neatly through the neck of 25 mm x 10 mm screw cap tubes
and w~ere autoclaved as a unit before use. With several hundred eggs in the
cylinder, the cylinder was dipped in the first tube containing 70% ethyl
alcohol for 15 seconds to facilitate wetting the egg surface (Hinton, 1968).
Thle cylinder was then transferred to a tube containing 105 Zephiran chloride
for 15 minutes w~ith occasional agitation. [Wdhite's solution (Whnite, 1931) for
15 minutes is also an effective germicide.] Eggs usually sank to the bottom
of the cylinder during immersion in thle germicide. Thle cylinder was next
transferred through 2 tubes of sterile wiater for 5 minutes each. An aliquot
Wluinthrop Laboratories, Nlew York, N.~Y.
broth. The culture flasks were incubated with a 12-hour light cycle in
an incubator. Pupation begin at the 8th day, and female pupa, recognized
by their larger head, were w~ithdrawin with a syringe fitted to a long 14 g
The gnotobiotic arthropod module
The gnotobiotic arthropod module or GAM (Figure II-1) is a sealed
lucite capsule allowing sugar feeding at one end, and blood feeding through
a heated artificial membrane at the other. Mosquito pupae were injected
through the rubber stopper (#3) with a small amount of water, which allowed
adult emergence and humidified the capsule when held at 27.C.
The GAM was constructed from .3 cm lucite cylinders measuring 7.5 cm
i.d. x 10 cm and cemented with PVC industrial solvent. About 2/3 of the
inner surface of the cylinder was covered with plastic screen to provide
resting sites. A Gelman~filter support (#5,9) was bonded to the GAM floor
(#8) and interposed a stretched gutta percha membrane (#r6). The outer
collar (#9) of this filter support was threaded so it can be inverted and
screwed tightly into the floor (#10) of the heating manifold. A tefloR
seal (#7) insured watertight integrity. The GAM wras sterilized writh the
membrane in place in an ethylene oxtide chamber, or autoclaved if the membrane
was to be inserted later (30 minutes at 15 1bs and 120"C).
Figure II-2 shows a battery of 4 GAMs in the blood feeding mode in
place in the heating manifold. The blood meal was injected through the
rubber stopper of the blood tube (#11) and was kept at a constant temperature
of 32oC. Heated w;ater was pumped from a laboratory water bath into the
manifold inlet (#3) causing an overflowJ (#2) return to the water bath.
Nutritional Biochemicals, Cleveland, Ohio
3Gelman Instrument Co., Ann Arbor, M~ich.
of 25-50 eggs was taken up in a small amount of water in a sterile pipette
and transferred into a sterile dry tube containing 1.5 cm x 10 cm strip of
filter paper. The eggs were dispersed over the str-ip in an amount of water
just sufficient to thoroughly wet the paper strip and -the tubes wJere incu-
bated on their sides at 27oC. Eggs hatched in 24 hours and the larvae were
seen crawling on the surface of the paper. Within the next 3 hours, the
tubes were filled witih sterile water, causing the empty egg shells to sink,
and the sterile larvae to float to the surface for recovery.
The larval diet
Formulation of the larval diet met the following requirements: good
growth and adult emergence after autoclave sterilization, standardization,
and ease of preparation. The composition of this diet is illustrated in
Table II-1. The RNA component and the casein wrere dissolved in 250 mil of
water over heat. Dilute KOH was added to maintain pH at about 7.0 as solu-
tion proceeds (Dadd et al., 1973). The hog supplements wias previously
pr-epared by drying at 140OF for 1 hour, ground in a mill, and sieved twice
through a 50 mesh sieve. The remainiing ingredients were added and final
volume brought up to 1 liter. The pH was adjusted to 7.0 with KOH and dis-
pensed to flat-sided 16 ounce screw~ cap prescription battles in volumes
sufficient to produce a depth of .5 cm: (about 50 m~is). Each culture bottle
contained a 5 cm x 15 cm strip of plastic screen miaterial for emerging adults
to rest. on when cultures w~ere continued to that stage. Finally, then culture
bottles were auto0claved for 15 minutes.
Newly-hatched larvae were inoculated 1 larva per 2 ml of mediaa using
a sterile Pasteur pipette. Sterili ty was moni tored by i nocul ati ng an
appropriate larval sample into tubes of thioglycollate2 broth and nutrient2
:Ralston Purina, St. Louis, Mlissouri
Difco Laboratories, Detroit, ilichigan
Gnotobiotic Arthropod Module (GAM) used for the Maintenance
andi Blood Feeding of Anopheles quadr~imaeulabuis: 1. Sugar
delivery tube and wick, 2. Cylinder top, 3. Rubber stoppers,
4. Lucite cylinder, 5. Outer collar of Gelman filter
support, 6. Gutta percha membrane, 7. Teflon seal,
8. Cylinder floor, 9. Inner Gelman filter support,
10. Floor of heating manifold, 11. Blood delivery tube.
Figure II-2. Battery of 4 GAbls in Position for Blood Feeding in Heating
Manifold: 1. Top of manifold, 2. Overflow into waterbath,
3. Inlet from waterbath, 4. Manifold tank.
Table II-1 Composition of Diet for Rearing
Alnophzelec quadr~-iimaea~atuIs Say
SIngredient mg/1,000 ml
Purina S.E. Hog Supplementl 40% 1,000
NBC Liver Powder2 500
Brewer's Yeast2 500
Ribonucleic Acid2 1,000
Vitamin Diet Fortification Mlix2 1,000
Salt Mixture W (Wesson)2 250
Ralston Purina Co., St. Louis, Missouri
Nutritional Biochemicals Corp., Cleveland, Ohio
Survival of mosquitoes in the GAMl was quite good provided leakage of blood
and sugar solution was prevented. Normally, 12-15 adult mosquitoes were
maintained in each GAM.
Several mlethodss for the surface sterilization of mosquito eggs wecre
tried including peracetic acid (Doll ~et al., 1963), antibiotics (Epps
et al., 1950), LysolR (Adkin and Bacot, 1917), C10roxR (Lea, 1957), iodine,
alcohol, ZephiranR (Dadd et al.,1973) and White's solution (W~hite, 1931).
Only the latter 2 germicides consistently produced viable sterile An4.
quadrlimaculabus lar-vae, and then, only when the eggs w~ere held on moist
filter paper. Effective egg sterilization deprived the egg of its ability
to float, and eggs which sink are non-viable. An extended embryonationn
period is another result of several of the germicides tested. Bacteria-free,
viable larvae could be consistently produced in 24 hours, using the sterili-
zation procedure described.
The larval diet
The larval diet was a composite of the diets suggested by several
workers. Liver powder and brewer's yeast1 are traditional protein sources
(Trager, 1935) which, in the present study, were included as first food for
first instar larvae. Purina S.E. hog supplement 40%2 is an excellent protein
source for older larvae and was used as the standard diet for conventional
Anopheles sp. Dadd et al. (1973) reported that ribonucleic acid, and
soluble casein, because it contained an essential steral as a contaminant,
1Difco Laboratories, Detroit, Michigan
Ralston-Purina, St. Louis, Missouri
were both critical components of their miosquito diet. For that reason,
both were included! in the present diet.
The 8 vitamnins essential to growth and development of mosquitoes
(Akov, 1962) were more than supplied by the addition of 1 gm/11ter NBC1
Vitamin Fortification Mixture. Although probably not utilized (Clements,
1963), it also contains vitamins A, 812, C. D, and E. Furthermore, it
is compounded in dextrose which, as Nayar (1966) and Johnson (1969) found,
contributed to maximum adult survival when included in Acdes sp. diets.
Originally, this vitamin component was sterilized by Millipore filtration,
but like Dadd et al. (1973), no adverse effect could be detected when the
vitamin component was autoclaved along wlith the rest of the media. Use of
the Saltlnixture-W2 very conveniently approximated Trager's (1935b) salt
mixture and evidently provides excellent buffering capacity in this appli-
Immoderately high osmotic pressure of the culture media can be re-
sponsible for poor growth and development on otherwise adequate chemically
defined media (Nayar, 1966). An. qu/adim-i~~aclluatus larvae are highly sen-
sitive to high osmotic pressure which can cause larval death in the first
25 hours of culture. For that reason, the diet described here was very
dilute in comparison to other diets (Rosales-Ronquillo et al., 1973).
Developmental time of germ-free mosquitoes was approximately the
same as for conventionally-reared mosquitoess. Approximately 75%/ of the
first stage larvae survived to theo pupal stage on the 8th day. N~o reduction
of vigor in pupae or adults was noted.
Nutritional Biochemicals, Cleveland, Ohio
MFillipore Corporation, Bedford, Mlass.
Use of the G~notobiotic Arthropod Miodule
The GAN provided an air-tight sterilizable capsule to maintain 12-15
gnotobiotic arthropods. Some practice was required in determining how much
water is injected with the pupae. In general, as little as possible will
suffice to humidify the capsule but will not entrap adults. Leakage of
sugar solution from the wick and blood seepage around the membrane are other
problems which may be encountered. Both can entrap mosquitoes if allowed
Considerable flexibility is possible with the GAM and it should pro-
vide appropriate experimental units for a variety of blood feeding arthro-
pods. The GAM may be successfully employed in comparative nutritional
studies of related arthropods, the identification of growth factors, trace
elements, and the effects of environmental variations. The influence of
contaminating microorganisms may be selectively eliminated for the study
of host metabolism, growth and aging under a variety of conditions.
Insect pathogen studies would be an appropriate application of the
GAM. Pathological material may be safely isolated, and its effects evaluated
without the influence of bacterial contamination. Pathological effects of
insect microsporidian parasites may be evaluated in the sterile intestine
and in conjunction with various combinations of indigenous microflora.
The GAM may have wide application in basic research on the relation-
ship between an arthropod host and the parasites it verters. Previous
gnotobiotic studies have shown that the success of many parasitic diseases
depends not only on the etiological agent, but also on the host's intestinal
microflora (Phillips and Gorstein, 1966; Mlicks and Ferguson, 1961;
Bradley et al., 1964). Parasites and potential parasites passing through
the arthropod host's alimentary tract are exposed to both the microorganisms
that reside there, and to the intestinal climate which mnay, in part, be
influenced by them. Bacteria present at the host-parasite interface mray
influence the ability or disability of the arthropod to vector parasitic
III. THE NATIVE MIDGUT BACTERIAL FLORA OF
SEVERAL WILD AND COLONIZED MOSQUITO SPECIES
The ability of mnosquitoes to vector disease is of such significance
to world-wide public health, that a better understanding of their bio-
logical relationships with microorganisms in general is critical. The
intimacy existing between the host and i ts mnidgut population of mnicro-
organisms suggests more than chance environmental contamination. In-
stead, years of conjunct evolution mnay have produced a highly integrated
gastrointestinal ecosystem which may ultimately participate in the ca-
pacity of an insect to transmit disease.
On the basis of a long series of compar-ative studies of albino mice
from different colonies, Dubos et al. (1965) have concluded that the
indigenous (or normal) microbiota of these animals exerts a profound
influence on their rate of growth, their efficiency in the utilization of
food, and their resistance to infection, toxic substances, and other stress-
ful agencies. Many attributes of miicE which are characteristic of the
colonies from which the animals were derived are in reality determined
not by genetic endowment, but by the microbiota prevailing in the colony.
There is no reason to suppose that insects should be unique among animals
in their association with microorganisms.
In the course of another study, it became of interest to define the
midgut bacterial population of wild filaria vectors. Two pairs of inde-
pendent wJorkers (Chao and W~istreich, 1959, 1960; W~istreich and Chao,
1960, 1961; Ferguson and Mick~s, 1961; Miicks et al., 1961; Miicks an~d
Ferguson, 1961, 1963), have produced a series of reports on colonized
mosquitoes which indicate that the flora of adults is different from
that of larvae; it varies among the different species, and it is
distinct from bacteria in the specimen's environment. Although there
w~as always a high percentage of sterile individuals, the microorganisms
recovered from contaminated specimens are generally non-pathogens widely
distributed in food, water, soil, and intestines of animals.
These reports, however, w~ere qualitative evaluations of laboratory
insects maintained under standard conditions in established colonies.
No reports have appeared which define the microbial composition of wild
naturally-feeding mosquitoess. The following report is a qualitative
and quantitative evaluation of several hundred adult specimens of
colon-ized Anophelec quladi-~nrcimatcult Say, wild An. quadimaiculabust~s wild
Anophl~c crucciians Wiedemrann,wild Aedes inf~irmaictu Dyar and Knrab, and
Al. erucians complex, trapped at 6 stations in Gainesville, Florida.
Materials and Methods
One-hundred adult femnale specimecns of An2. quadtEl~inc~ulatus (Gaines-
villo str-ain) were selected over a period of time froml the colonies at
the USDA's Agricultural Research Service, Insects Affectin~g Man Laboratory
in Gainesville, Florida. They had been fed a liverl:yeastl miixture
Nutritional Biochemicals, Cloveland, Ohio
for- the first 3 larval days and ground S.E. Hogl Supplementl for the
balance of the larval period. Adults were held 5-7 days after emer-
gence, fed a 2.5'; sugar solution, and starved 24 hours prior to dis-
Several hundred wiild mosquitoes w~ere trapped from 6 field stations
around Gainesville, Florida, with a CDC2 type light trap using dry ice
as a carbon dioxide source. The species composition of the wild females
selected for study w~as: An. qjuad~rimaculatus, 25 specimens; An?. orulcian,lc
80 specimens; Al. infi2rmatus, 12 specimens; An. orucian~s complex, 20
specimens. These mosquitoes were also held 24 hours without food before
Both oral and anal openings of each insect were sealed wyith aucoR
cement by bonding the insect at these points to a plastic cover slip.
Seepage of the germicide into the digestive tract was prevented, and
the coverslip effectively anchored the insect for dissection beneath the
surface of sterile water. The mounted insect was immersed in 70% ethanol
for 30-60 seconds, rinsed and dissected under sterile water. A section
of abdominal integument was removed and placed in a tube of thioglycollate
broth3 as a sterility check. Gentle distention of the lower abdomen ex-
posed the digestive tract, and the midgut wads retracted and removed w~ith-
out puncture. The midgut was then homogenized in .4 ml sterile water in
a glass tissue homogenizer. One-tenth ml aliquots of the homogenate
were then streaked for isolation onto duplicated! Eugon3 agar plates, 1 Eugon
Ralston Purina Co., St. Louis, Mo.
2 iiational Communicable Disease Center, Atlanta, Ga.
Difco Laboratories, Detroit, Mich.
agar plate fortified with 5% defibrina-ted sheep blood, and I tube of
fluid thioglycollatel. Incubation took place at 370C, 250C, 370C, and
370C, respectively. A total count of organisms present in the midgut
w~as made by counting all of the colony-forming units, appearing in
Eugon agar after 24 hours, and multiplying by a factor of 4. Each
different type of colony appearing on the 3 agar plates was counted,
sampled, assigned an isolate number and streaked for pure culture on
Eugon 1 agar. When growth appeared inl the thioglycollatel tube, ino-
culations were made into veal infusion agar tubes' to differentiate
obligate anaerobes. Yeasts and fungi were inoculated into Sabouraud
When pure cultures were apparent, inoculations w~ere made into triple
sugar iron agarl (TSI) and brain-heart infusion agarl (BHI) slants which
were maintained as stock cultures at 90C. Colony morphology, chromo-
genesis, and Gramn-reaction were determined at this time. Table 1
lists the procedure in the presumiptive determination of bacterial genera.
Hydrolytic abilities were determined in agar supplemented w~ith 5%:
soluble starch, 5% gelatin, 5% Wesson oil, or jjZ alphacell. Oxygen
relations were demonstrated in duplicate tubes (one overlaid with sterile
mineral oil) of OFI media and incubated f-or 5 days. The presumptive
identification of Gr'am-negative enteric bacteria as w;ell as the catalase
test were performed on TSI. Decarboxylase production w~as tested in
lysine, arginine, ornithine, and control tubes. An I.N~.V.i.C. series
(indol production, methyl red pH, Voges-Proskauer, and citrate utiliza-
tion) was performed in appropriate media. Both the nitrate reduction
Difco Laboratories Detroit, Mich.
test and motility were determined in nitrate agar1 slants. The type
of milk reaction w~as demonstrated in litmnus milk1 media anid the hydroly-
sis of urea was determined on urea agarl. Carbohydrate formentations
were demonstrated on cystine trypticase agar1 plates overlaid with
any 10 carbohydrate differentiation discs listed in Table 1.
Data analysis wras by means of a Statistical Analysis System (SAS)
program written by Barr and Goodnight (1972). A frequency distribution
of the species composition of midgut bacterial flora was computed.
Correlation coefficients and probability statements were calculated on
the frequency of occurrences of all pair combinations of bacterial isolates.
~Ades infirmatzue Dyar & Knab
The average midgut bacterial count of 8 specimens of A. infir-matis
was 263. No bacteriologically sterile specimens were encountered in
the sample. Table III-1 illustrates the species composition of midgut
bacterial flora isolated from all 5 groups of mosquitoes. The most
frequently isolated bacterial species in this mosquito group was Bacill~as
cereus Frankland and Frankland, non-pathogenic Coryniieba~ctrii- sp., anld
Micrococcu~s iuteus (Schroeter).
Ano~pheles quadr-ncimazculatus Say, (ild)
Sixty-one percent of the w~ild An,. quad~~rimculatuss examined w~ere
found to be bacteriologically sterile, which was the highest rate of
Difco Laboratories, Detroit, Mich.
Procedures in the Presumptive Determination of Bacterial Genera
1. Hydrolysis Reactions:
2. Oxidative Fermentation Reaction
3. Catalase Trest
4. Triple Sugar Iron Slant
5. Growth Study:
Eugon Agar @ 250C
Eugon~ Agar @ 37oC
6. Decarboxylase: Lysine
Nitrate Reduction Test
Litmus Milk Reaction
12. Fermentation Reactions (any 10):
~a n n itol
Spe~cis Co-iposition of Mijdgut Bacterlial Florii in 5 Groups of Mosquitoes
SPECIES CO;PO;ITIONI PERCENT OF TOTAL?)
c o. r
' ~ -; L~ ; L'-~:L
~ F 9
~iaS?~ n ~ r,
i-" ~ "" ~'~
t a ~~ N ~~ N
U _a ;? ~ ri
~ '" 6
` 4 .& 5
: ;i S .Z
SP;C;S C:L:.T h ~ u u o
"' ~ 3 B
~ v. .. i.
i.'.7 1 C3
ii, :5 3 ~2
3 37 16 4
" ~ B B ,,
i: i i. ~ r-j vl
09 a BS
r. ,, P, i
X S '~ ~ P .~
~ ~ $ ~ iS U
r 1: ~J it ?j C1
~Q ij o iI n t~ ~
'' ;; r;~ ~: X
i; u u r:
TABLE III-1 (continued)
SPECIES COMPOSITIONS PERCENTT OF TOTAL)
318 94 6
20 100 :
19 CO i "a 9~ Bf
__ ~I __
TArBLE III-1 (continued)
SPECIES CO;MFOSITION (PERCENT OF TOTAL.)
a ~ '''
3. i ~ a n .-!
q ~ ` ' ~? ~ ~
- jqi. -r :1
ii O C .I~ ;' ~ ~
u ~ t-
rl ~ U U C? U U u '-r w ii ~
, ~ ., N ii
u rj j ,., ~,
.. . h r. r:
i i ~ ii c : u 8
-.1 ., II i. ~~ -, 3
I -. -~ a g
c, ir x ,,
TABLE :iI-1 (continued)
SPECIES COMPOSITION PERCENT OF TOTAL)
o ;1 o
2 '" 'L
ii ~ d ~
ood~ a a
is ~ ~~ r~ ~ '' B
lr ') W
~4 i .i
lu f rl R
N (i e
~I ~1 F:
i u -~
~1 u :;
;~e~E III-I (continued)
SPECIES CGI1POS:TION I~ERCE~T OF TOTAL)
I: ? ei ~ a ~
e U '-' O '"
i) E3 .~
3 ii' c, ~ -I: c ~
h U ~ ~ .I ~ L1 ii 3 ii c
'li o ~ o r. g g n
Y~ 6~ ii
_:~ ~ 5 ~ U -: i Li V i.
c PJ ~ L3
'. ii ~ ~~ t : ~
;. u ~ L" C i LI
:1 X :
r- ~ r, ! ' -i i
. ~ tIO Y U ii :. U . r)
Tj-? i ii i, `1
~ i: ii '' ;. ~
' L ;; 2 ''' :: ''
I~ i ~; .
4 ~ Y
;':71 ;:.: Z-:
G;j i3 ? i9
TA8LE II-I- (continued)
SPECIES COMPOSITIONI PERCENT OF TOTAL)
ii :, ~-
a c, o
~ w L
ij , C L
1. ~I u
N j h II
C n og g ~YILI
~,.~ m ~
N C 3
o ~ c -I e
SI'Ec:Es ConT ;
calatu 13 100
continued) 100 5
-60 --5 5
15? 20i go
312i 96 $
TABLE III-1 (continued)
SPECIES COMPOSITIONI (PERCENT OF: TOTAL)
? '~J V)
C I c
Y i) ~ ?. I ^
~ ~ "
'' '* r
II ~. '' ~' '' ~ ~ L:
; or u i
-~ u u -i~ ~ -
L) I. L1 i.
F n ~-, ;,
rr r vi N ~
" d 1~
P3 :jl ~ ii rj i.
1 ~~` H E: E
*:::Ca ?;3 5 IC')
(colony) ;00 10
(con:tJnue) 90 100
TABLE iII-1 (continued)
SPECIES CODPOSITION (PERCENT OF TOTAL)
P, UU C
a S ;'"
~ "' D a 4 ~,.,
g ~ g ;j
i- L; o ~
,~ ~, o*IN ~g ,, ~,!
h ii 1 N
~ ~ B f~ u a
$on i a
( ~ % t-~ ir Lj C) i:
rlcs:u!rc! ~ ii
pi YI a
o r: y ~ e
u . i
C: s -
re .r ~ ; c
~ S g
d :: o
v i, u.a
- ~Y N O i~
B U '*
a U, .,
c. c~ :~ B ~,
o r, o
" g - 0
~-~ FI Y)
TkALE III-1 (continued)
SPECIES COMIPOSITION [PERCENT OF TCTAL)
a u N
F: B ,; ;~
ild "" I; -e
r.) m 1 n r
L~ aU r; , c_
YI ; L) U i, g ., Y1 YI ~ .I:
~ a u i~ rl n I .I u~
r~ u~g"~"S~' 3
n $ ~~ )J .I.i P C
C 1? i;
r, @ I
R " a : ` 6 ~
'" '5 F c
~ i: - i; s "
i~ X ~"
~1 T y 3 C:
TATI 3 ... i B
u r;; - 3
:ICSLI!TD eiL I~?:~L ,, ~ L r~
S'EZIES C~i"lT ii r; i~ P; U
rj ,: o X i,
r: ,~ s o
SN r~ N ~
~ ., .e ~ g
Q 3 :: ~'
<:4 c:ca 10 99
:aic: 2537 82 58
00O 1 9
2000 70 30
TABLE III-1 (continued)
SPECIES COMPiOSITION PERCENTT OF TOTAL)
Li .. L 7: 1--.
g il~ 'P L~ i.
'i C ri ii vi ri v r-
r? o V " 1'~ 'O Y~
* i. C
~rryJ~~e;m; ~? "
gs c` "3
4 a c! ~, P r:
P ~ '' 3 :'
ii i '~ '~ ':
X 9 ,
~~ c? u
no io O.;q. 'c,
't-, c, 'i
_i i 9 - ? r
"' ~ ~~ G ~ '" '"
e~ i: g X
N - E L1 '" ~ ~~ ~i ~ B
,, .:~"38"? E:
':""' LL ;-
midgut sterility encountered. Twenty-five specimens however, had an
average midgut bacterial counit of 666. The most frequently isolated
bacterial species were Enterobactel cloacaet (Jordan) followed by
^nopheles cr~c-ia a compl ex
Forty-three percent of this mosquito group were bacteriologically
sterile. The average midgut bacterial count of 25 specimens was 272
which was usually composed of Sa~lmone.Lla sp., Alcatigenes fa~ecaic
Castellani and Chalmers, and E. cloncae.
AF.nopheles crucian~a Wiiedemann
The midgut sterility rate of An. crucians wdas 36%. Eighty specimens
however, had an average midgut bacterial count of 377. The most fre-
quently isolated bacterial types were B. cereus, Salmonella sp., A.
faecal-ie, and Klebsietta. pnzemioniae (Schroeter).
Anophzieles quadrimacul~abus Say (colony)
Colonized An. quadrimaculatus had a midgut sterility rate of 40%.
Sixty individuals had an average midgut bacterial count of 214, which
w~as composed of B. cereus and the yeast, SaEccharmyices sp.
The Normal Flora
Sa7,mone.lla sp. appeared most frequently in An~. c~ruciana7 and Anr.
arucians complex. They produced very small (1-2 mm) mnucoid, discrete
colonies on Eugon agar and blood. A negative Gram stain demonstrated
m~otile, non-encapsulated rods. Miost isolates were catalase positive
and all were able to utilize citrate. Decarboxylase reactions were
variable, but most isolates carboxylated lysine and ornithine, and
produced H2S on TSI. Carbohydrate fermentations w~ere quite variable.
Antigenic typing of Sal~monella sp. was not attempted.
The Actinomycetales wdere represented by a group of non-pathogenic
Corynebacteriu~m and the genus Arthroba-"ter. The former were frequently
isolated from A. infirmatus~, while the latter appeared only in a
single isolate from a wild Ani. qu~1ad~~rim~aculatus The Coryn~jebacterfuml~
sp. produced grey to white star-shaped colonies on Eugon agar and
frequently produced luxuriant growth on blood agar without hemolyzing
it. These isolates were characterized by their high degree of pleo-
morphism which usually took the form of a non-spore forming Gram-
positive rod of irregular outline. Most isolates were catalase positive
and produced strong oxidative reactions on OF media. These isolates
were presumed to be non-pathogenic coryneforms because of their ap-
parently non-pathogenic nature.
A. faecalis w~as isolated most frequently fr~om Aln. oruccians. They
produced pale yellow, wrinkled colonies on Eugon's agar. They usually
occurred as motile coccoid Gram-negative rods. Mlost isolated produced
no change on TSI slants and negative reactions on indole, H2S and urea
sl ants. Carbohydrate fermentations were also negative. Simmons-citrate
reactions were variable, but all isolates produced oxidative type OF
E. e loacae wa s i sol ated from nea rly 50% of wilid An.? quad-rimacul~t atu~s
examined. These Gramn-negative motile rods usually produced smooth,
round, rose-pigmented colonies on Eugon agar. They wiere distinctive
in producing an acid TSI, positive citrate and a positive VP test.
Most isolates produced moderate to strong hydrolytic reactions on
starch agar, lipid agar, and gelatin agar. An occasional strain hydro-
lyzed cellulose (5% alphace0)~. Catalase reactions were positive and
most carbohydrates were actively fermented. nitrates were reduced by
B. cercus, was the domiinant midgut bacterial species of 95% of the
colonial Ain. qualdrirnaculaitus. that we!re contaminated. It appeared oc-
casionally in all other mosquito species. B. cereusE produced rough,
white lobate or round colonies on Eugon' agar. Their long, thick
rods usually occurred in long chains and produced cylindrical spores.
All isolates were catalase positive and produced strong hydrolytic reac-
tions on starch agar and gelatin agar. Decarboxylase activity was
usually not present. Carbohydrate fermentations were variable, but
arabinose, xylose, and mannitol typically w~ere not fermented. B. cereus
was differentiated from Bacilluts suibtilisi (Ehrenberg) which was occa-
sionally isolated from colonized An. q7uadr'imaclatu!Z 2s by their inability
to form chains and their characteristic litmus milk reaction.
KC. pn~eumonriseo was isolated most frequently fr~om Anz. crucians. They
produced 1-2 mm brown colonies on Eugon agar and appeared as non-
mnotile capsulated rods which occur singly or in short chains. Occasionally,
the fimbriae wJere recognized. This group typically produced negative
reactions in ornithine, and positive reactions in gelatin.Teypoud
acid and gas from glucose, utilized citrate, and produced a positive VP'
reaction. These isolate; often produced luxuriant growth with a green
non-hemolytic sheen on blood agar.
Nutritional BiochemicalsCo., Cleveland, Ohio
Mlicrocovenel! ro0selS Flugge and M, Zubea~si were isolated most fre-
qluently from A. iLafirma7tcs and An?. criiiucias, respectively. They are
Gram-positive cocci that produced large, opaque growth on Eugon's
agar under aerobic conditions. Cell division occurred in 2 or more
planes forming irregular clusters. Both groups were strongly catalase
positive and produced acid without gas from glucose. M~. roaeus. hydro)-
lyzed starch and gelatin, and produced bright red colonies on lipid
agar. Smooth, convex rose-colored colonies were produced on Eugon'
agar. Metabolism was of the respiratory type and decarboxylose activity
was negative. M~. luteus produced yellow, glossy colonies on Eugon
agar and was distinguished by its ability to hydrolize Fats. An un-
identified non-pigmented Miciococcus sp. was occasionally isolated from
wiild An. quladrima~culatuls which failed to hydrolize starch and gelatin,
and produced variably positive decarboxylase reactions.
PSEUdZomIonaS aerugin2osa (Schroeter) and Pseudao~n-nas au~rofaciene
Kluyver were isolated in low frequency from A. infirmnal lue and An. crucianis,
respectively. An unidentified Psuedoonanss sp. was isolated 3 times from
A. cru~cians. This family of bacteria are short, motile, Gram-negative rods
that are highly pigmented. P. aureofacienss produced bright orange dis-
crete colonies while P". a.-ruginosa developed red pigment. Neither species
gave a positive MR, VP, or H2S reaction. Negative decarboxylase reac-
tions occurred in lysine, ornithine, and arginine medias. Both species
gave alkaline milk; reactions and produced strong catalase refactions.
Nitrate reduction was variable.
Se'rat-ia mancsconzs Bizio w~as isolated from wild An. c,ucJians, complex.
These bacteria were motile, Gram-negative rods that produced smooth 2-3
mm colonies on Eugon agar. They gave strong catalase reactions and
reflected a greenish-i ridescenrce in oblique light. Most isolated pro-
duced brigh red colonies on lipid agar- and hydrolized starch agar.
They produced positive VP reactions and wer-e able to utilize citrate.
Metabolism was of thle fermentative type.
Es-cher-ich~ia co~i- (Migula) was isolated frequently from Anl. orucians
and An. oru~cians complex. These Gram-negative rods produced acid and
gas but no H2S on TSI. Nlo capsules or fimlbriae were noted, but isolates
were strongly catalase positive. Ornithiine, lysine, and arginine are
decarboxylated. Citrate was not utilized but acid and gas w~ere produced
from xylose, dextrose, maltose, lactose, and mannitol. Antigenic typing
was not performed.
Frequent isolates of Staphylococcus caprlophytiou~s (Fairbrother) were
made from An. crucia~s.. These Gram-positive cocci occurred in clusters
and produced smoothh, white convex colonies on Eugon agar. They were
facultatively anaerobic w~ith strong catalase reactions and variable de-
carboxylase activity. Acid was produced from glucose, lactose, and
sucrose. Nitrates wdere reduced and gelatin, lipids, and urea wdere
hydrolyzed. Citrate could be utilized as a sole source of carbon.
Streptcc~culs fan:calis Andrew~s and Harper and an unidentified Strepto-
coc.cu~s sp. wlere isolated fromn An. ecuicianz. These Gram-positive cocci
typically occurred in pairs or in short chains. Metabolism was of the
fermentative type alnd catalase tests w~ere negative. Colonies appeared
as discrete, white and muccid on Eugon agar and luxuriant on blood agar.
S'. faecalis did not hemnolyze blood, but the unidentified Streptococaves
produced strong beta he-molysis. An acid curd w~ith reduction appeared in
litmus milk and nitrates were not reduced. Both citrate and glucose
were actively fermented. Both cultures grew in the presence of .04%
tellurite. Antigen typing was not accomplished.
An unidentified isolate showing the characteristics of L~ctoacrcilZ1o
sp. 'was consistently isolated from An. cru~cians complex. These Gram-
positive, non spore-forming rods or coccobacilli form minute colorless
colonies on Eugon agar. They were consistently catalase negative
and failed to reduce nitrates. Indole, H2S, acetyl methylcarbinal, and
urea were not produced. They produced arginine dehydrolase but did not
split lysine or ornithine. Proteolysis was often produced in litmus
milk. Carbohydrate fermentation was variable.
Acinetobacter catcoaceticuis (Beijerinck) was isolated consistently,
but in low numbers from all 5 mosquito groups. These short, thiick Gram-
negative rods produced discrete smooth mucoid colonies on Eugon agar
and brown hemolytic colonies on blood agar. These isolates produced
a pr~oteolytic reaction in litmus milk, were catalase positive, and failed
to produce indole, H2S, reduce nitrate or liquify gelatin. All isolates
were oxidase negative and demonstrated weak OF reactions.
Three isolates from Ain. c2rucinans were tentatively identified as a
species of soil Niaisseria. These Gram-negative cocci w~ere aerobic anrd
formed large, yellow~ colonies on Eugon agar and were hemolytic on blood.
Thely wEre catalase and oxidase positive, and they produced indole, arrdacety
mefthylcarbinal. They u-tilized citrate as a sole carbon source but failed
to decarboxylate lysine, ornithine or arginine. Glucose and maltose
werle weakly attacked.
Species R P2
Neisseria sp.; SalmnellcZa sp. 0.334 .0002
Azomonas sp. ; S. facealis 0.272 .001
P. aureofaciens; A. faecalis 0.468 .0001
S. faecalis; A. faecalis 0.215 .001
Al. faecealis; M. iZutea 0.214 .012
S. Zu~tca; Salmonella sp. 0.379 .0001
A. faecalis; P. aeru2ginosa 0.177 .037
Ei. cereus; Salmonlella sp. -0.199 .001
B. eereus; A. faecalic -0.264 .002
Br. cereus; E'. clZoaeae -0.173 .004
B. cereus; K. pnewoioniae -0.180 .003
Correlation Coefficients of the Frequency
of Occurrence of Midgut Bacterial Species
The remain-ing bacterial species consisted of lowJ frequency isolations
or solitary isolations. These bacteria were Protceus vu~lgari~s Hauser,
Flavobacter~~-oiw ilutc~enis (Migula), Cibtrobacter frau;ndlii (Braak,), and un-
identified species of Planococcus, Alzobachce7ium, anld Amomr~ona. The yeast
Saccheromyces sp. was identified by morphological characteristics and its
ability to grow on Sabouraud maltose agar.
Correlation in Bacterial Occurrence
Table III-2 represents correlation coefficients and their probability
statements calculated for each combination of bacterial species in the
midgut flora. Salmnonella sp. is positively correlated (P>.001) with both
Micrococuis Zu.teuts and the unidentified Neisser-ia sp. Stre~ptoocous
fe7ecaZis is positively correlated (P>.001) with Azohaobersp. and A.
faecais although these combinations occurred in only 3 insects. A.
faea~is is positively correlated with M~. Zuteus- (P>.001) and P. aerugin~osa
The negatively correlated bacterial pairs all involved B. cereus.
b'. cercuo is negatively correlated with Sa~lmoneZZa sp. (P>.001), Al.
f'aecais (P>.001), E. cloacae (P>.005) anld P. pneumon~iae (P>.005).
The large proportion of bacteriologically sterile individuals is
not surprising. Chac and Wdistreich (1959) reported a sterility rate of
8% and 23%, respectively, for adult female and male Cu:~ex tar;aia
Coquillette. Wiistrich and Chao (1963) reported a midgut ster-ility rate
of 24% and 40Z, respectively, for female and male adult naeds iaogypti L.
Mlicks et al. (1961) reported mnidgut sterility rates of 50:5 for An!. quadri-
macuZatus females, 91%: for Al. ategy!pti, and 78% and 89X, respectively,
for Culex rat-igansz W~iedemamand cuLZey~ motestus2! Forskal. M~icks and
Ferguson (1963) in a later investigation, reported the midguts of all
C. molestus females examined wer-e uniformly sterile.
The presence of a bactericidal and/or bacteriostatic substance
in the midgut of certain mosquito species has some support. Duncan
(1926) demonstrated the presence of a bactericidal substance present in
the gut contents and feces of Aledes cincreu~s M~eigen and Anlopheles bi-
furcatuis (L) which was effective against 3 species of Ba~cillu*s. More
recently, the activity of antibacterial factors present in other blood-
sucking arthropods has been largely associated with and limited to Gram-
positive bacteria (Duncan, 1926, Anigstein et al., 1950).
The fact that microorganisms could not be isolated from a number of
the specimens in this and other studies should not be taken to mean that
microorganisms were not there. The standard cultural procedures em-
ployed in these studies are not adequate to demonstrate intracellular or
exquisitely fastidious organisms.
The species composition of the microflora isolated from the S groups
of mosquitoes in this study compares favorably w~ith other work~ers. The
isolates from adult C. tarsatis (Chao and Wistreich, 1959) showed the
presence of Achromobacter gutatus (Zimlmerman). Achromoajcter sp., E.
cloacae, c~itrobseber. intrmiedia (Braak), Fl~avobatterriumn sp., M~icroolccus
sp., M~icrco ccusf vivir-anzs M~igula, ProbetaLs sp., and Pseudomona~s sp. Cul-
tures of unidentified Gr~am-negative rods as well as SaccharomycesR sp. and
Geotirichu:n sp. wer~e also reported. The same investigators, Wistreichh and
Chao (1960), isolated bacteria from midgu-t sections of 14 fourth-instar
larvae of 0. tarsais that had been raised in fresh tap waiter and fed
brewer's yeast. Results of these isolations showed the presence of
Achrom~obancter sp., Tsactobacittlus sp., M.IC.'OsocL'uS CanizZd'I Cohn, M~icro-
coccuis sp., pr'oteus resttgel't Hadley et a., and P~seudomoarcr, sp. Again,
cultures of unidentified Gram-nega~tive rods and Saxchchromyces sp. and
Geotrichum; sp. were observed. Isolations from the larval rearing water
showed the presence of E. cloaczae, Kr~ebsiella sp., Sa~cchaxramuce sp. and
Geotr-ichum7 sp. Fourth instar larvae were found to contain examples of
all the genera of bacteria isolated from adult C. tarsai3 w~ith the
exception of Fnierobacter, Citrobatcter, and Flavobacterium,, whereas iso-
lations from adults did not show~ the presence of Lactobacillus or the
coral pigmented form of Micrococa.!
Bacteria isolated from the midguts of adult CuZez quinqjuefasciatus
Say fed apple slices for 2 days after emergence showed the presence of
Achromobac-ter sp., Flavobac~teriumi sp., KlecBielZra sp., Microcccurs cazseo-
lyticus Evans, and Poeudomnras sp. (Chao and W~istreich, 1960). M~idgut
sections from 11 fourth instar larvae of the same species produced only
Baociltis sp. and the yeast Sacchiammryces sp.
Bacteriologic evaluation of the moidguts of larval anid adult Aedes
sicervensic Ludlowi and Anzophece a'bimanus Wiedemannfed on a larval diet
of yeast and adult diet of apple slices was reported by Wlistreich and
Chao (1961). Larval isolation f~rom Al. albimanus showed the presence of
Alcaligenea sp., Bac~%i~ll sp. anid unidentified Gram-negative rodls. Adult
Al. alb~imian~u: showed Ah ~'iomoacter's~p., Alcal~igrene sp., and Corynebabcte,1:iu
sp. Sacchurequesli~r was isolated from both larvae and adults. A. sierrancsis
larvae contained RalciZue~ cimulous:: Jordan, and P~arac-colobactkum, intermedi
Bor-man. Alc~iacroba:cter sp. and c. itr~er~media were found in adults. Sac-
charomyeces sp. were isolated only from the larvae. Wlistreich and Chao
(1961) also repor-ted that mlicroorganisms isolated from the adult and
larval environments were distinctly different from isolates of the
Wlistreich and Chao (1963) reported bacteriologic evaluations on the
midguts of larval and adult A. aegypti which had been fed yeast and
apple slices, respectively. Isolations from larvae showed the presence
of the following bacteria: Alcaiigeneis sp., Cory~nsbceba-Cteiu sp., c.
freundii, urbhialz beesonii Hauduroy et al., M. luteus, as well as fungi
belonging to thle general Geotlichumv, Peniiciiuim, and Saccharomyces.
Isolations from adult mosquitoes showed the presence of Achromobacter
sp., AZealigenzes sp., and Saccharomyce~s sp.
Ferguson and Micks (1961) reported additional experiments using
adult female C. fatigans from a colony that had been denied access to
food. By means of a sterile micro-dissection technique, they demon-
strated the presence of 7,actobacitus sp., Aiealigeness sp., Pseuidomoza~s
sp., and 2 unidentified cultures of Gram-negative rods. Bacterial iso-
lates from the midgut of similarly tr-eated adult female An1. quaSdlvima:cuSa~tus
proved to be 2 strains of Screpiococzus and a species of Aerobaeber (Micks
and Ferguson, 1963). Al. aegypti produced only a species of Colyneb:3cacterium
and no microorganisms could be cultured -from the m~idguts of adult female
Generally, these bacteria represent common soil, water, or enteric
or-ganisms widely distributed in nature. Four of the 6 collection sites
from which Anz. crucians~, Ain. cr~caians complex and w~ild An. quladrimacul~atus
were collected w~ere locations around a large lake on the University of
Florida campus. This lake receives a number of sewe~rage effluent. The
presence of such enteric organisms as P. vi'lga1'ri, C.
fas~cais in the mnidguts of detritis feeders is not surprising.
Colonial mosquitoes tend to exhibit a microFlor-a made up of 1 or
2 bacterial types which are consistent for the colony. WJild mosque toes
usually have a variety of microorganisms which are generally comparable
within the species.
A notable exception proved to be the isolation of Net~s~cori sp.
from 3 specimens of An. crudenc~. They r~esemnbled thle N~icrococcaceae.
somewhat, but bore an even closer resemblance to 2 types of Ieisneria
described in grasshoppers by Blucher andi Stephens (1959). One isolate
was tentatively identified as N. catalrshalic (Frosch and Kolle). The
taxonomic position of these 2 types is uncertain and it is questionable
if they belong to this family or if they should be considered a member
of thle M:icrococcaceae that had lost their ability to retain the Gram
stain or to ferment carbohydrates.
The significance of the positively correlated bacterial combinations
is difficult to interpret. S. ~faecais and Ai. farccaicci might be ex-
pected to occur together because? of their enteric origin. The other
combinations mlay be the result of mutually satisfying metabolic relation-
Significance of the negatively correlated comibinat-ions favoring
thle presence of P. comm~u is moore obvious. The family Cacillaceae produce
a variety of extracellular products. Those products secreted by B. cerease
include hemnolysin, soluble toxin lethal for mice, enzymles lytic for
bacterial cells, proteolytic enzymies and phospholipase C. (Buchanan
and Gibbons, 1974). R. cerlius has been incrimninated in food poisoning
where extensive multiplication has occurred in foods. By wiay of its
secreted products, this bacterial species apparently competitively
inhibits most other intestinal microorganrisms, w~ith th~e yeast Sacchzaro-
myces being a notable exception.
IV. EFFECTS OF BACTERIOLOGICAL
FLORA ON THE EARLY DEVELOPMENT OF
DIROFILARIA IMITIlrS LEIDY IN
ANOPELE QUA\DRIMACULATUS SAY
The relationship of the normal intestinal microflora to the vectoring
capacity of mosquitoes has not been adequately investigated. Insects are
not unique among animals in their relationships with microor-ganisms.
Parasites and potential parasites, passing through the arthropod hosts
alimentary tract are exposed to both the microorgansism that reside there,
and to the intestinal climate which may, in part, be mediated by them.
Indigenous bacterial flora present at the host-parasite interface must be
regarded w~ith as much consequence as the host itself.
Dirofitarias inn7i;tic Leidy, the etiological agent of canine heartworm
disease, requires passage through thle mosquito to the L3 or infective
stage. Although susceptible or resistant mosquitoes species are genetically
determined reviewedd by MacDonald, 1973), the physiologic mechanism
whereby the insect accepts or rejects the infection has not been elucidated.
The presence or absence of host micr-oflora, or the species composition of
micr~oflora, mnay be one of the factors mefdiating the establishment of D.
ilmmithi in susceptible mnosqu itoes.
Gnotobiotic studies have shown that the success of many parasitic
diseases depends not only on the etialogic agent, but also upon the host's
intestinal mnicroflora. Phillips and Gorstein (1966) reported that the
association of the amioeba Entamoeba ~h~i a-totytica Kruse) with one of 3
species of bacteria was necessary to produce typical hepatic abcesses
in hamsters. Phillips (1964) considered the bacteria acted by pro-
viding a physically and chemically suitable environment favorable to
survival of the amoeba, but W~it-tner and Rosenbaum (1970) proposed the
existence of some critical factor passing from bacteria to pathogen.
Mlicks and Ferguson (1961) concluded that microorganisms play a signifi-
cant role in the susceptibility of mosquitoes to malarial parasites.
They suggest either a competition between the microorganism and ma-
larial parasite for essential nutrients, or that certain strains of
microorganisms supply factors needed by the host for metabolic processes
and defense mechanism~s. Elimination o-f the microorganisms in the latter
case would presumable increase host susceptibility to malaria.
This paper reports the possible effects of midgut bacterial flora
isolated from 5 groups of wild and colony mosquitoes on the ability of
gnotabiotic mosquitoes to support development of first stage D. iinitis
Materials and Methods
Eggs of Alnophelces quair~imnaculatis~ Say were obtained from the colonies
of the USDA Agricultural Research Service, Insects Affecting Man Labora-
tories in Gainesville, Florida. Egg sterilization procedures, larval diet,
and germ-free rearing techniques have been previously described (Results
II). Eggs w~ere surface sterilized in 10%J Zephiran; allowed to hatch,
and the sterile larvae injected into 16 ounce prescription bottle rearing