NATURAL-BASED REPELLENT PRODUCTS: EFFICACY FOR MILITARY

AND GENERAL PUBLIC USES

Author(s): ROBERT J. NOVAK and EUGENE J. GERBERG
Source: Journal of the American Mosquito Control Association, 21(sp1):7-11.
Published By: The American Mosquito Control Association
DOI: http://dx.doi.org/10.2987/8756-971X(2005)21[7:NRPEFM]2.0.CO;2
URL: http://www.bioone.org/doi/full/10.2987/8756-971X%282005%2921%5B7%3ANRPEFM
%5D2.0.CO%3B2

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and
environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published
by nonprofit societies, associations, museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of
BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries
or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research
libraries, and research funders in the common goal of maximizing access to critical research.
Journal of the American Mosquito Control Association, 21(4) Supplement:7–11, 2005
Copyright q 2005 by the American Mosquito Control Association, Inc.

NATURAL-BASED REPELLENT PRODUCTS: EFFICACY FOR

MILITARY AND GENERAL PUBLIC USES
ROBERT J. NOVAK AND EUGENE J. GERBERG

INTRODUCTION AD), a Greek physician who described how worm-

The first use of repellents as a means to avoid wood (Artemisia absinthium) could repel gnats and
the bites of noxious mosquitoes and other biting fleas.
arthropods was probably done by prehistoric man Oils extracted from plants or plant parts with a
when they found that sitting next to a smoke-gen- variety of volatile compounds (Tables 1 and 2) that
erating fire tended to provide a degree of protec- repelled insects were the principal means that peo-
tion. This method is still used in many places by ple used for protecting themselves or their domestic
remote peoples and at times by modern man when animals prior to World War I, which was the advent
camping out in natural areas. Today, even the most of synthetic chemical repellents. Oil of citronella
effective insect- or mosquito-abatement programs was one of the most widely used early repellents,
cannot totally eliminate the nuisances caused by being introduced around 1882. It was registered in
these blood-seeking arthropods (Novak and Lamp- the USA as McKessons Oil of Citronella and was
man 2001). Therefore it is necessary at times to used for human application to repel both mosqui-
employ personal protection to help avoid insect toes and gnats. Other essential oils that were used
bites. There are a wide variety of methods that peo- during this time were anise, bay laurel, bergamot,
ple can employ to help avoid noxious insects by cassia, cedar wood, eucalyptus, and wintergreen.
establishing a mechanical barrier, like bed-nets, In 1929, dimethyl phthalate, one of the first syn-
screened windows, and clothing. In the U.S.A., and thetic repellents, was reported to be effective
to a large degree in Europe and Japan, the primary against house and stable flies. Grannett (1940) was
barriers are hermetically sealed and air-conditioned instrumental in developing ethyl hexandiol, also
houses. Although clothing can be used to help know as Rutgers 612. Grannett also developed re-
avoid unwanted insect bites, such as loose fitting producible methods to evaluate and test repellent
long-sleeved shirts and pants, the public tends to compounds. In 1937, indalone, dimethyl carbate
expose as much skin as possible during the hot, was synthesized at Rutgers University and reported
humid summer months, when these insects are at to be an effective repellent.
their highest densities. Clothing can, in fact, be a During World War II, it became necessary to de-
very practical and efficient way to avoid these velop effective and long-lasting insect repellents,
pests. primarily mosquito repellents, to help military
The objectives of this article are to define and troops avoid those biting insects that transmit path-
discuss the differences between natural or herbal- ogens causing debilitating diseases such as malaria,
based repellents with synthetic chemical-based re- dengue, and typhus. Just prior to World War II, the
pellents. In addition, we will discuss how to deter- U.S. Department of Agriculture established a lab-
mine efficacy of repellent activity using cage- and oratory at Orlando, FL, which was later moved to
field-testing protocols. We will also present a brief Gainesville, FL, which tested thousands of potential
discussion of those repellent methods that do not repellent compounds. During the war, the standard
work, especially those that have recently inundated military repellent was 622. This repellent contained
the market place. 6 parts dimethyl phthalate and 2 parts each of ethyl
hexandiol and indalone. After World War II, the
standard military repellent applied to the skin of
BRIEF HISTORY OF REPELLENTS soldiers was called M-2020, which was a mixture
Prior to World War II, a number of plant prod- of 40% dimethyl phthalate, 30% ethyl hexandiol,
ucts, generally oils and other substances derived and 30% dimethyl carbate. The standard repellent
from specific plant species, were the mainstay re- applied to clothing was M-1960, a mixture of 30%
pellents, applied to either skin or clothing to avoid benzyl benzoate, 30% n-butylacetanilide, 30% 2-
biting arthropods. The first methods man used to butyl-2-ethyl-1,3propanediol, and 10% TWEEN
repel insects was with smoke, covering the skin 80.
with mud, or by applying a variety of animal fats Diethyltoluamide (DEET), developed by Mc-
and greases. Native Americans rubbed cedar tree Cabe in 1954, is one of the very best mosquito
needles or used the fat from bear to prevent the repellents developed (Rutledge et al. 1996, Elston
nuisances caused by insects. Some wore various 1998) for mosquito control. It has been the gold
herbs and poultices around their necks or on their standard for public health threats and emergencies
clothing. The first written records of repellents as well as for use in military operations. During the
against insects occur in classical Roman literature recent outbreak of West Nile virus in the USA,
by Pliny, the naturalist, and by Dioscorides (40–90 DEET is the repellent of choice and has been part

7
8 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 21, NO. 4

Table 1. A brief list of natural or herbal products, by name, what they are derived from, the active ingredient or
ingredients, and insects for which it has been shown to have repellent activity.
Name Derived from Active ingredient Species
Citronella oil Cymbopogon (grass) Terpines, alcohols, aldehydes Biting flies
Geraniol coeur Geraniol Geraniol, nerol, citronellol Biting flies
Quwenling Lemon eucalyptus p-Methane-3, 8-diol Mosquitos
Neem Neem tree Many compounds Biting flies
Soy beam oil Soy bean plant Lecithin base, many compounds Biting flies
Thujic acid Western cedar Diethylamide Aeges aegypti
Tarweed extract Hemizonia sp. 1, 8-Cineole (eucalyptol) Ae. aegypti
Palmarosa oil Cymbopogon martinii Many compounds Anopheles sp.

of the standard recommendation for general public on the literature and they are presented merely as
use by the Centers for Disease Control and numer- a reference to the variety of natural products
ous state public health departments. Military forces claimed to have repellency. In general, natural
or public health workers exposed to vector-borne products are considerably less effective than syn-
diseases must have the means to protect themselves thetic repellents, like DEET, at equivalent concen-
for long periods of time. In addition, insect pest trations. Several chemicals and essential oils do
populations may exist in such large numbers that provide relatively short periods of mosquito repel-
they may interfere with military operations, requir- lency (Rutledge and Gupta 1996).
ing a long-lasting and durable repellent. There are several newly formulated repellents
Military operations and public health emergen- that have recently come to the market and may il-
cies normally do not occur in backyards of suburbia lustrate the future trend in repellent research. Merck
or at other outdoor recreational areas, where shorter Company has shown that 3-(N-butylacetamino)-
duration and less harsh repellents may be satisfac- propionate, called Merck 3535, has strong repellent
tory. In recent years, there has been a tendency to- action against biting flies and mosquitoes (Marchio
ward the use of natural agents or phytochemicals 1996). KBR 3023, a Bayer product, which is a pi-
(Tables 1 and 2). Some of the so-called natural perdine derivative (piperidinecarboxylic acid, 2-(2-
products, or, better described, herbal products are hydroxyethyl)1-methylpropylester), is also very ef-
listed in Table 2. Those compounds that have fective in repelling biting flies and mosquitoes. A
shown specific repellency to insects are briefly de- compound called PMD is similar to quwenling, but
scribed in Table 1. is derived by a novel extraction process. The active
A complete list of phytochemicals and essential ingredient is p-methane-3, 8-diol combined with is-
oils reported to have repellent activity toward vec- opulegol and citronella. PMD has been shown to
tors and other arthropods is beyond the scope of provide 5 h complete protection when tested
this review. The list (Table 2) contains chemicals against Anopheles funestus and Anopheles gambiae.
that have been reported in the literature as repel- Though permethrin, a pyrethroid, is not technically
lents, although, in some cases, their efficacy is un- a repellent, it is used in a manner similar to repel-
supported (e.g., the use of Vitamin B1). We have lents. Permethrins are used to impregnate military
made no attempt to rate their effectiveness based uniforms, clothing, and bed nets.

Table 2. A partial list of natural products and essential AGENTS: NATURAL VERSUS CHEMICAL
plant oils with repellent properties. With the advent of more natural-product repel-
Natural products identified lents, it is important to define the specific nature of
Limonene Vitamin B1 Cineole the repellent action. First, many potential natural
Linalool Geraniol Quwenling repellents can be considered a barrier to the insect,
Camphor Rotundial Thurjic acid preventing either landing or penetration of the skin.
Eucamolol In many cases, these barrier compounds can be skin
Plant essential oils lotions or sun screens. However, it is important that
Neem oil Garlic oil Birchwood tar these barrier repellents are field tested with the
Aniseed oil Thyme oil Soybean oil same rigor as volatile repellents in order to estab-
Geranium oil Eucalyptus oil Nutmeg oil lish product efficacy. Volatile compounds rely on
Bergamot oil Pyrethrum Pine oil vapor pressure and temperature to release specific
Lavender oil Coconut oil Clove oil chemicals that insects avoid. In fact, the operating
Orange-blossom oil Penny-royal oil Cinnamon oil definition of a volatile repellent is that its ‘‘Efficacy
Peppermint oil Citronella oil Palmarosa oil of a repellent product is based upon the vaporous
Castor seed oil Lemon oil Indian privet stage of the repellent to prevent bites.’’ In other
Tarweed
words, a minimum vapor pressure is required to
SUPPLEMENT 2005 CURRENT TRENDS IN REPELLENT TECHNOLOGY 9

maintain the efficacy of a repellent. Most repellent When a natural product- or herbal-based repel-
products on the market today use various delivery lent is developed, it is important to take into con-
mechanisms that either control the evaporation rate sideration that it will be used by the lay public for
of a repellent for long-term repellency or as the nuisance management, not for military or public
carrier agent disintegrates it releases repellent va- health emergencies. Military and public health
pors to prevent bites. This is a highly desirable workers use repellents in a different manner and
characteristic of most current and future repellent under completely different conditions than those
active ingredients. Based on the operating defini- that the general public should or would encounter.
tion, one of the key differences between these 2 In fact, Gupta and Rutledge (1994) state that the
types of repellents is the experimental design crit- use of repellents to reduce human/vector contact
ical to establish efficacy. With barrier repellents, an and reduce the transmission of mosquito-borne dis-
insect may land but not be able to bite. This change eases has not been scientifically proven. The 1998
in behavioral capability must be noted in the ex- U.S. Center for Disease Control and Prevention
perimental design because a land has a different (CDC) report concluded that human disease can be
meaning when testing a volatile-based repellent. effectively reduced by active surveillance systems
One of the other important differences between and appropriate mosquito-control measures, but re-
chemical- and natural product-based repellents is pellents are not mentioned as a specific means of
the duration of repelling activity. This question has prevention. Several factors concerning the general
been somewhat resolved by the Environmental Pro- public have to be taken into consideration when
tection Agency (EPA) by setting a 2-h minimum developing and testing a natural repellent product.
activity requirement of a repellent in order to gain The most effective insect repellent will not provide
registration. It is well established that several for- a uniform level of protection to all users under all
mulations of DEET can give up to 6 h complete conditions. Individual factors that can affect the de-
protection. To date, there is no natural product that gree of protection afforded by the repellent are 1)
can give that duration of control. However, due to Does the person apply the repellent properly? 2)
potential human-safety factors, it is recommended Individual susceptibility to mosquito bites, 3) per-
that DEET-containing products not be reapplied af- spiration, 4) personal product that may enhance or
ter the first application. Also, it is not recommended detract from the repellents effectiveness, and 5) en-
that DEET be combined with skin lotions or sun vironmental factors, such as wind, temperature, hu-
screens because these products are generally reap- midity, mosquito species, age, and density. Because
plied by the user. This is the big difference between repellents are personal protection, the application is
DEET and natural products, which can be reapplied totally dependant on the user and, unless an indi-
safely, thus compensating for the shorter duration vidual actually dips him or herself in the repellent,
of repellent activity. complete body coverage cannot be guaranteed.
The number of insects biting plays a major role Also there is a great deal of variation among indi-
in not only testing a repellent but in the marketing viduals on the effectiveness of a repellant, thus
of a product. Most marketing commercials tend to ‘‘one size does not fit all.’’
use the extremes in terms of insect population num-
bers. In fact, in most of these advertisements, the
insect numbers greatly exceed what the typical lay DETERMINATION OF EFFICACY
person would ever encounter or would remain in Cage studies
those circumstances. The recommendation of the
EPA’s Scientific Advisory Panel (SAP) for biting In regard to the utility of cage studies to test
pressures appropriate for testing should be based repellent efficacy, the findings of the EPA SAP
primarily on what the general public perceives as a should be followed. The SAP strongly recommends
nuisance problem. Because little information is that only field studies should be used to establish
available in the literature, experience of members efficacy and registration. Cage studies are not a val-
of the SAP coupled with a publication by Morris id substitute for repellent field studies. Cage tests
and Clanton (1988) titled ‘‘Quantification of a nui- should be used only as a screening device and
sance mosquito problem in Florida’’ were used as should not be submitted in support of a registration.
guidelines. It is important to remember that repel- They should, however, be used by the manufacturer
lents are used almost wholly for nuisance problems, to screen possible repellents, develop formulations,
not disease problems. Therefore, it follows that the and determine a range of application rates. The
guidelines used regarding biting pressure reflect Klun and Debboun (2000) device may be an alter-
conditions that impact the general public and not native to the device specified in the American So-
military or public health personnel. Based on these ciety for Testing and Materials (ASTM) standard
parameters, the SAP recommended the following for laboratory studies of mosquitoes (ASTM 2000;
biting rates for field testing insects and thus estab- 951-94), remembering that it is still a screening tool
lishing minimum nuisance thresholds of: mosqui- that was never intended as a substitute for field
toes 1 bite per min; ceratopogonids at 1 bite per 5 studies. However, if a test cage with an enclosed
min; tabanids at 1 bite per 5 min. area, such as Klun and Debboun (2000), that does
10 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 21, NO. 4

not provide for free flow of repellent vapors from establish efficacy and registration. Cage studies are
the surface and eventual dissipation of repellent va- not a valid substitute for repellent field studies.
pors into the immediate environment, it is probable Cage tests should be used only as a screening de-
that some repellents may have erroneously indicat- vice and should not be submitted in support of a
ed higher repellency. Any laboratory test cage se- registration. They should, however, be used by the
lected for product testing should take vaporous manufacturer to screen possible repellents, develop
state of repellents into account before being rec- formulations, and determine a range of application
ommended for use. rates. The Klun and Debboun device may be an
alternative to the device specified in the ASTM
Application amounts (dosage) standard for laboratory studies of mosquitoes
(ASTM 2000; 951-94), remembering that it is still
The amount of repellent to be applied to the skin a screening tool that was never intended as a sub-
should be determined by the registrant or manufac- stitute for mosquito field studies. It is very impor-
turer. There are several reasons that justify this tant to remember that any laboratory test cage se-
statement. The amount applied is determined by lected for product testing should take into account
weight, which makes it very difficult to determine the vapor state of repellents before use.
application rates of aerosols. Therefore, because
most repellents are liquids, creams, or aerosols, the
application rates should be in milliliters or in sec- Biting frequency
onds of spray time for aerosol. Also, the test area The primary goal for testing any product is to
for application of 600 cm2 is too large an area for employ good science to assure validity of the re-
many arms. A test area of 250–300 cm2 is more sults. The purpose for testing the performance of
than appropriate. repellants is to establish a product’s capability of
The amount of the repellent to be tested should preventing pest arthropods from generally annoy-
be determined by the registrant. This could be de- ing, puncturing the skin, or taking a blood meal
termined by scientifically conducting statistically from humans. Historically, the EPA has used the
valid studies that demonstrate the quantity of a given first confirmed bite test to assess the effectiveness
physical formulation consumers are likely to apply. of human insect repellents. However, the concern
Apparently, some of these data already exist in the of the EPA and the scientific community is that the
cosmetic industry. Where they do not exist, repellent first confirmed bite method will result in the loss
manufacturers should conduct them. EPA could, and of valuable data. The first confirmed bite method
probably should, serve as a repository for this in- does not appear to have been developed using a
formation. The dose rate per unit could then be es- statistically valid approach, whereas alternative
tablished through prefield tests using cage tests. The methods, such as first bite or a 95% reduction, pro-
rationale for this is that there are and, in the future, vide a statistically valid real-world assessment of
will be numerous new products that do not fit the insect-repellent efficacy. The 95% reduction in
synthetic chemical repellent mode of action. We are bites requires a study design that allows compari-
already seeing this with the number of natural re- sons between a control and treatment. This requires
pellents and many new products that have multiple that a control be used each time a new treatment
purposes, i.e., sun screen, moisturizers, etc. This (or set of treatments) is studied. This would provide
would certainly play a major role in determining the for a statistically generated testing protocol, which
application amount. The EPA could then use field- would also give EPA standardize statistically ana-
efficacy data for registration and labeling. lyzed data for comparing both skin-applied and bar-
rier repellents. It would also take into account var-
Field studies iations in test subjects, location, and product dose,
The current resurgence for natural- or herbal- and/or formulations. Moreover, in a short period of
based repellents has caused a re-examination of time, EPA would have a comprehensive database
how repellents should be tested in order to establish on numerous natural and chemical repellents,
efficacy and future U.S. EPA registration. Several which could certainly aid in the registration pro-
key issues will be discussed regarding field testing, cess. The null hypothesis would also be standard
including: 1) cage versus field studies, 2) establish- for all products and, of course, based on 95% biting
ing biting frequency, 3) experimental design, and inhibition. This type of testing design would also
4) need for a Good Laboratory Practice (GLP) field eliminate the controversy about complete protection
protocol. The following discussion will be based in time (CPT), first confirmed bite (FCB) because
part using information from the U.S. EPA’s Scien- 95% reduction would be the standard. Also, the
tific Advisor Panel on repellents. question concerning the need for a GLP protocol
would become a mute point because the experi-
mental design would dictate the parameters of the
Cage versus field studies test. CPT for repellents would than be reduced to a
The authors, as well as the SAP, strongly rec- defined period of time to include 2 h. Two hours
ommended that only field studies should be used to should be the standard minimum time required for
SUPPLEMENT 2005 CURRENT TRENDS IN REPELLENT TECHNOLOGY 11

repellent activity. Employing the 95% reduction in add anything to the science and quality of the test
biting within the scope of the experimental design except additional costs.
would certainly provide a statistically sound pro-
tocol to test synthetic repellents as well as those SUMMARY
derived from natural or herbal-based products.
The major points addressed in this article regard-
ing natural- or herbal-based repellents are as fol-
Experimental design lows. 1) It is important to recognize the end user
when developing and testing repellents. There is a
One of the major problems with field studies re- major difference between repellents suitable for
volves around sample size or the number of human nuisance management by the general public as
test subjects required for a statistically valid test. It compared with military or public health uses when
is essential that sample size is based on the scien- battle-field conditions or a disease risk is present.
tific experimental design and not on formula-driven 2) The experimental design used to determine re-
guidelines. Gupta and Rutledge (1999) point out pellent efficacy must also recognize the end user.
that there are inherent flaws in the determination of Pest densities should reflect real-world conditions
sample size. For example, in reality, according to as much as possible. 3) The experimental design
Rutledge and Gupta (1999), for 5 individuals, the should be directed to the pest population, not at a
confidence of protection is 97.5% for 1 h but, at 2 different product. Remember, we should use sci-
h, it is only about 50%. Additionally, using protec- entific logic, not market-driven tactics. 4) There is
tion periods of 1–8 h with confidence limits of 99% a need for specific regulatory standards for natural
and 95%, the best possible results (P , 0.01 with repellents to insure not only quality but the proper
D 5 0.5 h) for a product claiming 1 h of protection and most effective means of application and use. 5)
would require 15 test subjects and one claiming 8 There is a need for increased research and devel-
h would require 280 individuals. This is not feasible opment for natural and synthetic repellents.
or practical. Using the 95% reduction in bite pro-
tocol, fairly standard experimental design with as- REFERENCES CITED
sociated variability information, and target confi- ASTM. 2000. E 939-94. Standard method of field testing
dence necessary, sample sizes should be quite easy topical applications of compounds as repellents for
to compute. In contrast, this would not be the case medically important and pest arthropods (including in-
if the first bite or the first confirmed bite protocols sects, ticks, and mites): 1 Mosquitoes.
were used. Elston DM. 1998. Insect repellents: an overview. J Am
In any experimental design, it is the case that the Acad Dermatol 36:644–645.
number of replicates of product(s) is of greater im- Grannett P. 1940. Studies of mosquito repellents. 1. Test
portance than the number of subjects. The principal procedure and methods of evaluating test data. J Econ
issue is ensuring tests are replicated a sufficient Entomol 33:563–565.
Gupta RK, Rutledge LC. 1994. Role of repellents in vec-
number of times to strengthen their statistical pow- tor control and disease prevention. Am J Trop Med Hyg
er. For example, if you test 4 repellent concentra- 50(suppl):82–86.
tions, you should use 5 people (4 treatments 1 1 Klun JA, Debboun M. 2000. A new module for quanti-
control) per replicate. You would repeat the test at tative evaluation of repellent efficacy using human sub-
least 5 times by rotating the treatments/control jects. J Med Entomol 37:177–181.
among the 5 participants. However, it is important Marchio T. 1996. Insect repellent 3535. SOFWJ 122:480–
to recognize that there are a number of perfectly 485 (in German).
adequate experimental designs to evaluate repel- McCabe ET, Barthel WF, Gertler SI, Hall SA. 1954. Insect
repellents. III. N, N-Diethylamides. J Org Chem 19:
lents. It was recommended by the SAP that, rather
493–498.
than dictate a single inflexible protocol, it may Morris CD, and Clanton KB. 1988. Quantification of a
make more sense to convene an expert panel to nuisance mosquito problem in Florida. J Am Mosq Con-
evaluate proposed testing protocols. This panel of trol Assoc 4:377–379.
experts could evaluate proposed protocols, pro- Novak RJ, Lampman RL. 2001. Public health pesticides.
posed statistical analyses, and provide protocol(s) In: Kreiger R, ed. Handbook of pesticide toxicology
acceptable to the EPA for field testing repellents. Volume 1. London, United Kingdom: Academic Press.
p 181–201.
Rutledge LC, Gupta RK. 1996. Reanalysis of the CG
GLP field protocol MacNay mosquito repellent data. J Vector Ecology 21:
132–135.
The EPA should not require GLP standards for Rutledge LC, Gupta RK. 1999. Variation in the protection
field trials. GLP was designed for laboratory stud- periods of repellents on individual human subjects: an
analytical review. J Am Mosq Control Assoc 15:348–
ies. The L in GLP 5 laboratory. If the EPA adopts 355.
a scientifically based experimental design for field- Rutledge LC, Gupta RK, Mehr ZA, Buescher MD, Reif-
efficacy studies, the standards for laboratory studies enrath WG. 1996. Evaluation of the laboratory rabbit
under GLP would be not be necessary. The GLP model for screening topical mosquito repellents. J Am
standards, as currently used for field studies, do not Mosq Control Assoc 12:142–143.