• Apium graveolens;
  • celery;
  • Aedes aegypti;
  • DEET;
  • mosquito repellents;
  • commercial repellents;
  • laboratory repellent bioassay;
  • field repellent bioassay


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In our search for new bioactive products against mosquito vectors, we reported the slightly larvicidal and adulticidal potency, but remarkable repellency of Apium graveolens both in laboratory and field conditions. Repellency of the ethanolic preparation of hexane-extracted A. graveolens was, therefore, investigated and compared with those of 15 commercial mosquito repellents including the most widely used, DEET. Hexane-extracted A. graveolens showed a significant degree of repellency in a dose-dependent manner with vanillin added. Ethanolic A. graveolens formulations (10–25% with and without vanillin) provided 2–5 h protection against female Aedes aegypti. Repellency that derived from the most effective repellent, 25% of hexane-extracted A. graveolens with the addition of 5% vanillin, was comparable to the value obtained from 25% of DEET with 5% vanillin added. Moreover, commercial repellents, except formulations of DEET, showed lower repellency than that of A. graveolens extract. When applied on human skin under field conditions, the hexane-extracted A. graveolens plus 5% vanillin showed a strong repellent action against a wide range of mosquito species belonging to various genera. It had a protective effect against Aedes gardnerii, Aedes lineatopennis, Anopheles barbirostris, Armigeres subalbatus, Culex tritaeniorhynchus, Culex gelidus, Culex vishnui group and Mansonia uniformis. The hexane-extracted A. graveolens did not cause a burning sensation or dermal irritation when applied to human skin. No adverse effects were observed on the skin or other parts of the human volunteers’ body during 6 months of the study period or in the following 3 months, after which time observations ceased. Therefore, A. graveolens can be a potential candidate for use in the development of commercial repellents that may be an alternative to conventional synthetic chemicals, particularly in community vector control applications.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Mosquito-borne diseases, such as dengue haemorrhagic fever, encephalitis, malaria and filariasis remain a major source of illness and death worldwide, particularly in tropical and subtropical climates (Becker et al. 2003). Protection against insect bites is best achieved by avoiding infested habitats, wearing protective clothing and applying insect repellent (Curtis 1992; Fradin 2001a). Repellents that can be used anywhere at any time are the only feasible measure for preventing arthropod attack in some situations. Economical and practical insect repellents have, therefore, become a viable and attractive alternative that is widely sold on the market. Commercially available insect repellents can be divided into three categories, synthetic chemicals, botanicals and alternatives such as the combination of synthetic and natural compounds. The most widely marketed synthetic-based insect repellent is N, N-diethyl-3-methylbenzamide (DEET), which has been used worldwide since 1957. It is currently effective and available in commercial formulations, e.g., solutions, lotions, gels, creams, aerosol sprays, sticks or impregnated towelettes (Robbins & Cherniack 1986; Golenda et al. 1999). This compound, however, has an unpleasant smell, oily feel and high skin penetration, and it can dissolve plastics and synthetic rubber (Qiu et al. 1998). Many consumers are unwilling to apply DEET on their skin, and deliberately seek out alternative repellents. Numerous repellents of natural origin or synthetics combined with botanicals in various formulations are currently on the market. Therefore, it is time to search for effective alternatives to synthetic chemicals, compare them with insect repellents on the market, and make recommendations for their prevention against mosquito bites.

Apium graveolens L. (Umbelliferae), celery, is a common European plant that is now grown, consumed and used as a popular vegetable and spice all over the world. Due to its pleasant odor and nutritional, pharmacological and medicinal values (Bisset 1994; Rafikali & Muraleednaran 2001), this plant product is deemed to be suitable for humans, pets and the environment. The literature, however, offers few publications on the antimosquito potential of this plant (Rafikali et al. 2000; Rafikali & Muraleednaran 2001). Our previous study on the antimosquito potential of A. graveolens (Choochote et al. 2004; Tuetun et al. 2004) reported its slightly larvicidal and adulticidal potency against Aedes aegypti mosquitoes. Fortunately, it clearly demonstrated the remarkable potential of this plant product for repellency, both in laboratory and field conditions. In continuation of our studies, we undertook the repellency investigation of the ethanolic preparation of hexane-extracted A. graveolens compared with those of 15 commercial mosquito repellents and the standard repellent, DEET. Results obtained from this study can be used to compare and select a suitable repellent for personal protection against mosquito attack, thus minimizing disease transmission and nuisance bites.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Test repellents

The seed of celery, A. graveolens, was obtained from E.A.R. Samunpri, a commercial supplier in Chiang Mai province, Thailand. The voucher specimen named PARA-AP-001/2 was deposited at the Department of Parasitology, Faculty of Medicine, Chiang Mai University, Thailand. Dried and powdered seeds (2 kg) were successively extracted three times by maceration, with 6 l of hexane at room temperature for 7 days. After vacuum filtration through a Büchner funnel, the hexane extract produced was concentrated to dryness on a rotary evaporator at 60 °C, until the solvent completely evaporated, and then lyophilized to yield a dry hexane extract. Vanillin (EC No. 2044652) was purchased commercially from Sigma-Aldrich® (Switzerland). This vanillin was produced for Research and Development (R&D) use only, not for drug, household or other uses. Vanillin, one of the most widely used of aroma chemicals (Curtis & Williams 1994), is used in almost all fragrances as a base fragrance note and to slow down the evaporation process to make fragrances last longer on the skin. For repellency evaluation, hexane-extracted A. graveolens as well as DEET were prepared in various formulations: 10–25% in absolute ethanol with and without 5% vanillin.

Fifteen commercial repellents were purchased from various stores in Chiang Mai province. They were selected to represent a range of active ingredients that included synthetics, botanicals and a combination of synthetic and botanical compounds. Names, ingredients and formulations for each repellent are illustrated in Table 1.

Table 1.  Name, ingredient(s) and formulations of 15 commercial mosquito repellents
Name and ingredient(s)FormulationProtection time shown (h)
Synthetic chemicals
Off!: DEET (14.25%), S.C. Johnson & Son Inc., USASpray7
Kor Yor 15 DEET lotion: DEET (24%), dimethyl phthalate (24%), Bride Cosmetics, Co. Ltd., ThailandLotion7
Kor Yor 15 IR lotion: IR 3535 (12%), Bride Cosmetics Co. Ltd., ThailandLotion7
Kor Yor 15 IR gel: IR 3535 (10%), Bride Cosmetics Co. Ltd., ThailandGel7
Sketolene roll on: IR 3535 (10%), The British Dispensary (L.P.) Co. Ltd., ThailandRoll on6
Beauti spray: IR 3535 (20%), Era Creation Inc., ThailandSpray7
Citronella lotion: citronella oil (10%), Faculty of Pharmacy, Chiang Mai University, ThailandLotion3–4
Citronella oil: citronella oil (15%), Rungudomchai Co. Ltd., ThailandSpray
Noxy: citronella grass oil (15%), Saint Louis Co. Ltd., ThailandSpray4–6
Herbal mosquito repellant: citronella oil (100%), Doi Kham, ThailandSpray3–4
Tipskin: bergamot oil, citronella oil, camphor oil, vanillin, Asia and Pacific Cosmetics Inc., ThailandCream4
Combination of synthetic and botanical compounds
Mistine censor: IR 3535 (12%), rosemary, lavender, eucalyptus, International Laboratory Inc., ThailandLotion6 1/2
Soffell (Floral Fragrance): DEET (13%), geranium, P.T. Herlina Indah Inc., IndonesiaLotion7
Soffell (Fresh Fragrance): DEET (13%), orange, P.T. Herlina Indah Inc., IndonesiaLotion7
Sketolene lotion: DEET, Eucalyptus citriodora oil (15%), The British Dispensary (L.P.) Co. Ltd., ThailandLotion7

Test mosquitoes

Pathogen-free, laboratory-reared nulliparous A. aegypti female mosquitoes were available for the experiments. The laboratory colony was reared according to the standard protocol of Limsuwan et al. (1987), and maintained at 25–30 °C and 80–90% relative humidity under a photoperiod of 14:10 h (light/dark) in the insectarium of the Department of Parasitology, Faculty of Medicine, Chiang Mai University. Adults maintained in screened cages were provided continuous access to 10% sucrose and 10% multivitamin syrup. Rats were used as a source of blood meal for females during egg production. Prior to testing, female mosquitoes (5–7 days old) were starved by providing them only water for 8–12 h.

Human volunteers

Healthy adult volunteers of either sex (aged 16–38 years) were recruited from the students and staff of the Department of Parasitology, Faculty of Medicine, Chiang Mai University. After being interviewed, volunteers who had no history of dermatological disease and allergic reaction to arthropod bites, stings or repellents were instructed in the methodology, probable discomforts to subjects and remedial arrangements, before signing an informed consent form.

Laboratory repellent bioassay

Ethanolic A. graveolens formulations with and without 5% vanillin, were evaluated for repellency in comparison to 15 commercial mosquito repellents and the standard repellent, DEET, against laboratory-reared A. aegypti females by using the human-treated procedure of the WHO (1996) standard method, with slight modifications. Two hundred and fifty 5–7-day-old starved female mosquitoes were placed in a net mosquito cage (30 × 30 × 30 cm) and allowed to acclimatize for 1 h. As a day biter, experiments on A. aegypti were carried out between 08.00 and 16.00 h in a 10 × 10 × 3 m room at 25–30 °C and relative humidity of 60–80%. The forearms of the volunteers were covered by plastic sleeves with a rectangular cut out 3 × 10 cm and matched to the ventral part of the forearm, permitting skin exposure of only the curbed area, and the hands were protected by rubber gloves. The test sample was applied as a dose of 0.1 ml evenly on the 30 cm2 exposed area of one forearm of each volunteer, and allowed to dry for 1 min. The other forearm was treated with a solvent vehicle (absolute ethanol) or solution of 5% vanillin in ethanol by the same protocol as that for the test repellent and used as the control. Due to the standardized avidity of the mosquitoes, the control arm was placed into a test cage for 3 min, at 30-minute intervals, prior to each insertion of the treated arm. If at least two mosquitoes landed on the control arm, the repellency test was carried out by exposing the treated forearm in a similar manner. Between each testing period, the volunteers were asked to relax in a room and avoid touching or rubbing the test sites on their forearms. The exposure experiment was continued at 30-minute intervals until the test forearm received two or more bites in the same observation period, or until a first bite occurred, followed by a confirming bite (second bite) in the next observation period. The complete protection time was recorded as the time that elapsed between application of the test sample and completion of the observation period immediately following it, in which a confirmatory bite was obtained. Each sample was tested six times with an identical test on different days for each of the four human volunteers (two females, two males). No one tested more than one formulation per day. The order of tests was assigned randomly and the volunteers were blinded to the repellent used.

Field repellent bioassay

The best repellent formulations (with the longest-lasting protection time) of each group, A. graveolens, DEET and commercial products established from the laboratory bioassay, were the candidates tested in the field assessment. The field repellent bioassays were conducted at minimal risk in human use procedure (Choochote et al. 1999; Ansari et al. 2000) in an area at low risk from vector-borne diseases. Experiments were undertaken against a natural population of mosquitoes, from 18 to 31 October 2004 in Nong Chom subdistrict located on Highway 1001, which is 5 km west of San Sai district, Chiang Mai province. A small canal ran along one side of the study location, which consisted of scattered open areas surrounded by human dwellings, trees, bracken and many larval habitats producing large mosquito populations that were sufficient to evaluate repellency. Preliminary trials by human-bait landing catches (October 10–17, 2004) revealed that some species of Anopheles, Aedes, Armigeres, Culex and Mansonia were abundant, whereas, A. aegypti mosquitoes were infrequently collected. Sunset at the test site occurred at ≈18.00 h local time during the study period. The volunteers were exposed to natural field populations of mosquitoes for 120 min, between 18.00 and 20.00 h, because preliminary surveys indicated that mosquito biting commenced 10–30 min after sunset and then decreased markedly after 90 min.

The treatment in the field study consisted of three formulations, A. graveolens (25 + 5% vanillin), DEET (25 + 5% vanillin) and Soffell lotion, and a control (solution of 5% vanillin in ethanol). Eight volunteers were divided into two groups comprising two females and two males (testers: 3, control: 1) for each group. For each trial, the test sample was applied as evenly as possible between the knee and ankle on the volunteers of each group. At the field site, the two groups of volunteers were situated at least 20 m apart. Three testers and a control subject of each group were invited to sit on chairs spaced 5 m apart and both legs were exposed for 120 min comprising twelve 10-minute periods. Mosquitoes landing on the legs of treated and control volunteers were captured by experienced mosquito collectors, with the help of a mouth aspirator and flashlight, and transferred into separate cups to be counted and identified later. After each 10-minute period, the volunteers moved to a new site at least 10 m from the last one. Volunteers and collectors were instructed to avoid applying any cosmetic products including perfume, cologne or lotion on the day of trial. Untreated areas on the volunteers were protected from mosquito bites by wearing a jacket with hood, gloves, socks and long trousers folded up to the knee. Collectors, subjects and their positions were rotated nightly, thus compensating for any factors affecting estimates of the repellent's efficacy such as position and personal differences, which included the number of mosquitoes, catching ability, skin absorption and persistence of repellent and attractiveness to mosquitoes.

The species of mosquitoes collected were identified under a stereomicroscope using the taxonomic keys of Tanaka et al. (1979) and Rattanarithikul and Panthusiri (1994). This protocol permitted us to determine which mosquito species were biting, total number of bites during the exposure period, biting rate and percentage repellency provided by the test samples, as compared with the control.

Data management and statistical analysis

The median complete-protection time was used as a standard repellency measure of the test samples against the A. aegypti mosquito in the laboratory bioassay. Differences in significance were determined by comparing the range of protection time in each sample. The effect of vanillin in prolonging the protection time of the repellents, ethanolic preparations of A. graveolens and DEET, was analyzed using the Mann–Whitney U-Test. The total number of mosquitoes collected during each exposure in the field trial was log-transformed before means and 95% CI were analyzed. A Kruskal–Wallis one-way anova was used to determine the significance of difference between the controls and those volunteers treated at the critical level of 0.05, using SPSS version 10.1. Percentage repellency (% Repellency) in the field bioassay was analyzed according to the following formula (Sharma & Ansari 1994; Yap et al. 1998):

  • image

where C is the number of mosquitoes collected from the control's legs and T is the number collected from the treated legs.

Ethical approval

The repellent protocol was reviewed and approved by the Research Ethics Committee of the Faculty of Medicine, Chiang Mai University, Thailand.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Laboratory repellency

The hexane extract of A. graveolens seed, once lyophilized, was a semi-solid non-polar material with a light-brown colour, pleasant fresh orange-like odour and yield of 5.03 % (w/w). According to the results presented in Table 2, it appeared that repellents tested under laboratory conditions possessed repellency against A. aegypti in varying degrees, and the most effective was the highest concentration of DEETv (25 + 5% vanillin), as measured by median protection time (5.5 h). Ethanol preparations of A. graveolens and DEET showed a significant degree of repellency in a dose-dependent manner with vanillin added. Increasing the active ingredients, A. graveolens or DEET, from 10% to 25% prolonged the protection time, which ranged from 2 to 3.5 h. Additionally, the protection time of these formulations increased significantly by the incorporation of 5% vanillin (P < 0.05) in all cases. Commercial repellents exhibited a wide variance in protection periods that ranged from 0 to 4 h, according to their active ingredients. All of the synthetic-based repellents (1–3.5 h) and the combination of the synthetic and botanical compounds (1–4 h) provided significantly better protection than those of botanical repellents (0–0.5 h). The best repellent formulations in each group, A. graveolens, DEET and commercial products assessed from the laboratory bioassay including 25% AHEv, 25% DEETv and Soffell lotion, respectively, were selected for further study under field conditions.

Table 2.  Repellency of various formulations of A. graveolens and DEET with and without 5% vanillin, and 15 commercial products against A. aegypti female mosquitoes under laboratory conditions
RepellantsMedian complete- protection time (range, h)
A. graveolens hexane extract: AHE
 10% AHE2 (1.5–2)
 15% AHE2 (1.5–2)
 20% AHE2 (2–2.5)
 25% AHE3.5 (3–3.5)
AHE + 5% vanillin: AHEv
 10% AHEv3 (3–3.5)
 15% AHEv4.5 (4–5)
 20% AHEv4.5 (4.5–5.5)
 25% AHEv5 (5–6)
 10% DEET2 (1.5–2)
 15% DEET3 (2.5–3)
 20% DEET3 (2.5–3.5)
 25% DEET3.5 (3.5–4)
DEET + 5% vanillin: DEETv
 10% DEETv4.5 (4–4.5)
 15% DEETv4.5 (4.5–5)
 20% DEETv5 (4.5–5)
 25% DEETv5.5 (4.5–5.5)
Commercial products
 Off!: DEET (14.25%)3.5 (3.5–4)
 Kor Yor 15 DEET lotion: DEET (24%), dimethyl phthalate (24%)3 (2.5–3.5)
 Kor Yor 15 IR lotion: IR 3535 (12%)2 (1.5–2)
 Kor Yor 15 IR gel: IR 3535 (10%)1.5 (1.5–1.5)
 Sketolene roll on: IR 3535 (10%)1 (1–1.5)
 Beauti spray: IR 3535 (20%)2 (1.5–2)
 Citronella lotion: citronella oil (10%)0 (0–0.5)
 Citronella oil: citronella oil (15%)0.5 (0–0.5)
 Noxy: citronella grass oil (15%)0 (0–0.5)
 Herbal mosquito repellant: citronella oil (100%)0 (0–0.5)
 Tipskin: bergamot oil, citronella oil, camphor oil and vanillin0 (0–0.5)
Combination of synthetic and botanical compounds
 Mistine censor: IR 3535 (12%), rosemary, lavender and eucalyptus1 (0.5–1)
 Soffell (Floral Fragrance): DEET (13%), geranium4 (3.5–4)
 Soffell (Fresh Fragrance): DEET (13%), orange4 (3.5–4)
 Sketolene lotion: DEET, Eucalyptus citriodora oil (15%)3 (2.5–3)

Field repellency

Tables 3 and 4 summarize the results of three selected repellent products, 25% AHEv, 25% DEETv and Soffell lotion, when applied to human skin under field conditions. These formulations showed excellent repellency against a natural mosquito population. There was a highly significant difference between the mean number of bites on the control subject and volunteers treated with 25% AHEv, 25% DEETv and Soffell lotion (Table 3). Only one bite from Anopheles barbirostris was observed in the same volunteers treated with 25% AHEv or 25% DEETv, whereas those volunteers treated with Soffell received none. This clearly demonstrated that 25% AHEv and 25% DEETv was effective in reducing bites with 99.9% protection and the protective effect of Sofell appeared complete (100% protection). Additionally, all products yielded excellent personal protection against a wide range of mosquito species belonging to five predominant genera, Aedes, Anopheles, Armigeres, Culex and Mansonia. Species and the number of mosquitoes caught are demonstrated in Table 4. In a total of 1252 adult mosquitoes comprising 15 species, A. barbirostris was the most predominant and accounted for about 32.7% of all the mosquitoes collected.

Table 3.  Mosquito biting rate on the legs of human volunteers, after treatment with solution of 5% vanillin in ethanol (control), 25% AHEv, 25% DEETv and Soffell lotion against field populations of mosquitoes
TreatmentTotal no. of mosquitoes collectedMosquito/10 min* (95% CI)
  1. A. barbirostris.

  2. * Mean in each column for each sample followed by the same symbol (‡, §) is not significantly different (P > 0.05).

Control12505.5‡ (4.7, 6.2)
25% AHEv   1†
25% DEETv   1†
Soffell lotion0
Table 4.  Results of field testing in three formulations, 25% AHEv, 25% DEETv and Soffell lotion, showing the protection offered against the various species of mosquitoes collected
Mosquito speciesControl25% AHEv25% DEETvSoffell lotion (commercial product)
No. of mosquitoes collected (%)No. of mosquitoes collected (%)% ProtectionNo. of mosquitoes collected (%)% ProtectionNo. of mosquitoes collected (%)% Protection
  1. ND, Not determined, as few specimens of this species were collected.

A. gardnerii131 (10.4)0 (0)1000 (0)1000 (0)100
A. lineatopennis11 (0.9)0 (0)1000 (0)1000 (0)100
A. albopictus1 (0.08)0 (0)ND0 (0)ND0 (0)ND
A. barbirostris410 (32.7)1 (0.08)99.81 (0.08)99.80 (0)100
A. peditaeniatus8 (0.64)0 (0)ND0 (0)ND0 (0)ND
A. tesellatus3 (0.24)0 (0)ND0 (0)ND0 (0)ND
A. jeyporiensis2 (0.16)0 (0)ND0 (0)ND0 (0)ND
A. sinensis1 (0.08)0 (0)ND0 (0)ND0 (0)ND
A. subalbatus357 (28.5)0 (0)1000 (0)1000 (0)100
C. tritaeniorhynchus74 (5.9)0 (0)1000 (0)1000 (0)100
C. gelidus70 (5.6)0 (0)1000 (0)1000 (0)100
C. vishnui group18 (1.4)0 (0)1000 (0)1000 (0)100
C. quinquefasciatus2 (0.16)0 (0)ND0 (0)ND0 (0)ND
M. uniformis160 (12.8)0 (0)1000 (0)1000 (0)100
M. indiana2 (0.16)0 (0)ND0 (0)ND0 (0)ND
Total1250 (99.8)1 (0.08)99.91 (0.08)99.90 (0)100


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The results of protection time extended by the addition of vanillin, which enhances repellency by lowering the evaporation rate of repellent from the skin surface, corresponded to those of Khan et al. (1975) and Tawatsin et al. (2001). Median protection times afforded by A. graveolens alone were slightly lower than those of DEET, when compared on an equal concentration basis. With vanillin added, however, A. graveolens products provided improved protection, which was comparable to that of DEETv. This was probably due to the lower evaporation rate and better skin persistence of A. graveolens, after the addition of vanillin. Our findings, along with those of earlier studies, suggested that A. graveolens extract should be formulated with vanillin for improving repellent efficacy.

All botanical products tested in this study consisted of citronella oil, as one of the active ingredients. Surprisingly, protection times that derived from these herbal repellents were very low (<1 h). Those of four products (Citronella lotion, Noxy, Herbal mosquito repellant and Tipskin) were less than 30 min and they were considered non-repellent. Although concentrations of active ingredients likely affected the efficacy of each repellent, some formulations with a lower dose of DEET, such as Soffell (13%) and Off! (14.25%), were found to be more effective than those of Kor Yor 15 DEET lotion (24%). Similarly, many products that have an equal amount of active ingredients such IR 3535 exhibited a different degree of repellency. The constituents of Mistine Censor, except IR 3535 (12%), are botanicals that include rosemary, lavender and eucalyptus. However, the repellent efficacy of Mistine Censor became significantly lower than that of Kor Yor 15 IR lotion (IR 3535: 12%). This clearly showed that repellent activity of each product depended on not only the dose of active ingredients, but also other synergists or additives added in each formulation. The protection time of Off! (3.5–4 h) obtained from this study was lower than that reported by Barnard and Xue (2004), when tested against Aedes albopictus (5.5–8 h). Differences in these results could be attributed to the variety of techniques used to measure mosquito repellency, the mosquito species tested and experimental environment. In addition to those differences previously described, a lower efficacy than that shown on the containers of all commercial repellents tested in this study may be partially influenced by the stability of each product, particularly the plant-based repellents. However, it is pleasing that all formulations of A. graveolens yielded a repellency of equal or more than 2 h, which is the minimum protection time required for registration at the FDA for marketing. The apparent results gave weight to the possibility of developing A. graveolens-based products as a potentially commercial repellent to supplant or replace conventional synthetics.

From results of its avidity and convenience as a laboratory mosquito, A. aegypti is extensively used worldwide for repellent screening, and its sensitivity to active ingredients can be an indicator of repellent activity (Badolo et al. 2004). Nevertheless, the protective effects against A. aegypti under laboratory conditions of a candidate repellent may not ensure success against other mosquito species under similar or different circumstances, particularly in the field. In our field investigation, the selected repellents, 25% AHEv, 25% DEETv and Soffell, afforded encouragingly good personal protection against a wide range of mosquito species. The most predominant of those species collected was A. barbirostris, which was also the only one to land on volunteers treated with 25% AHEv or 25% DEETv during the peak time for mosquito biting. These findings indicated that, in addition to mosquito susceptibility to each repellent, the density of biting mosquitoes present in any given area likely affected the apparent repellent activity. Although Soffell provided slightly better protection than 25% AHEv and 25% DEETv, the values of the mean number of bites and % protection obtained from these repellents showed no statistically significant difference. With regard to our results, the repellency of three selected repellent products after 2 h of application was approximately equal. Additionally, their protection against Armigeres subalbatus, a vicious biter and the other mosquito vectors, Aedes gardnerii, Aedes lineatopennis, Culex tritaeniorhynchus, Culex gelidus, Culex vishnui group and Mansonia uniformis appeared complete. When compared with other field studies, the complete protective effect of A. graveolens against A. subalbatus was significantly higher than those of neem oil (37.5%) and the ethanol-extracted Curcuma aromatica (89.69%) tested in the studies of Sharma et al. (1995) and Pitasawat et al. (2003), respectively. The three candidate repellents may possibly protect against other mosquito species including A. albopictus, Anopheles peditaeniatus, Anopheles tesellatus, Anopheles jeyporiensis, Anopheles sinensis, Culex quinquefasciatus and Mansonia indiana, which were collected in numbers too low to allow a valid estimation of the level of protection against them. During the 6-month period of the study or in the following 3 months, after which time observations ceased, no adverse effects such as irritation, rash, dermatitis or other allergic responses were observed on the skin of the volunteers who applied these repellents. Some volunteers, however, complained about an unpleasant smell, oily feeling and a hot sensation from repellents containing DEET, both in ethanol preparations and commercial formulations. With the exception of an oily feeling after applying A. graveolens at the highest dose (25%), formulations of this application caused no noticeably undesirable effects, although allergic responses such as urticaria and dermatitis in some celery farmworkers or food processors have been reported (Palumbo & Lynn 1953; Pauli et al. 1985). Nevertheless, this study may not adequately reflect the effects of long-term and routine use of celery extracts as topical repellents, during which time skin irritation may develop.

Results obtained from this study conducted under laboratory and field conditions suggested that three products, which comprised 25% AHEv, 25% DEETv and Soffell lotion (DEET: active ingredient), yielded approximately equal repellency, and were potential repellents for use as personal protection against mosquito attack. All DEET-based products that we tested under laboratory conditions offered a remarkable protection, ranging from a median of about 2 to 5.5 h. DEET, an effective broad-spectrum repellent, has been used for over 40 years and nearly 8 billion human applications, and there are relatively few documents of adverse effects (Idem 1998). DEET-based repellents can be expected to provide a safe as well as long-lasting repellency, when applied according to instructions. However, they are not perfect repellents because DEET may be washed off by perspiration or rain, and its efficacy decreases dramatically with rising outdoor temperatures (Maibach et al. 1974a,b; Fradin 2001b). DEET is also a plasticizer, capable of damaging watch crystals, eyeglass frames, and certain synthetic fabrics (Qiu et al. 1998; Fradin & Day 2002). Because of the unpleasant odour and uncomfortable oily or sticky application on the skin of some DEET formulations, many customers are reluctant to apply them. Some prefer to use plant-derived repellents, particularly ingested compounds, which nowadays seem to be more acceptable alternatives. User acceptance is one of the important factors affecting the potential use of repellents. Due to the limitation of an oily feeling when A. graveolens is applied in a high dose and its historically allergic responses, further research on long-term toxicity to human health and the appropriate formulation in reducing unpleasant feelings and improving safety for consumers is necessary. Reformulation to improve repellent potency and stability is also needed. It is worthwhile to develop and produce a plant-based repellent that can qualify for registration as a commercial product, which may be an alternative to conventional synthetic chemicals, particularly in community vector control applications. As this plant is cheap and available in many different parts of the tropics, the commercial exploitation of celery-based repellents would also contribute towards economic development, particularly in developing countries like Thailand, which is used to having economic crisis.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors sincerely thank the individuals who served as subject volunteers. Special thanks also go to the staff members of the Department of Parasitology, Faculty of Medicine for their cooperation. Financial support from the Thailand Research Fund through the Royal Golden Jubilee Ph.D. Program (Grant No. PHD/0112/2547) is acknowledged. Acknowledgment is extended to the Faculty of Medicine, Chiang Mai University for funding the publication cost.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Ansari MA, Vasudevan P, Tandon M & Razdan RK (2000) Larvicidal and mosquito repellent action of peppermint (Mentha piperita) oil. Bioresource Technology 71, 267271.
  • Badolo A, Ilboudo-Sanogo E, Ouédraogo AP & Costantini C (2004) Evaluation of the sensitivity of Aedes aegypti and Anopheles gambiae complex mosquitoes to two insect repellents: DEET and KBR 3023. Tropical Medicine and International Health 9, 330334.
  • Barnard DR & Xue RD (2004) Laboratory evaluation of mosquito repellents against Aedes albopictus, Culex nigripalpus, and Ochlerotatus triseriatus (Diptera: Culicidae). Journal of Medical Entomology 41, 726730.
  • Becker N, Petrić D, Zgomba M et al. (2003) Mosquitoes and their Control. Kluwer Academic/Plenum Publishers, New York.
  • Bisset NG (1994) Herbal Drugs and Phytophamaceuticals. Scientific Publishers, Stuttgart.
  • Choochote W, Kanjanapothi D, Panthong A et al. (1999) Larvicidal, adulticidal and repellent effects of Kaempferia galanga. Southeast Asian Journal of Tropical Medicine and Public Health 30, 470476.
  • Choochote W, Tuetun, B, Kanjanapothi D et al. (2004) Potential of crude seed extract of celery, Apium graveolens L., against the mosquito Aedes aegypti (L.) (Diptera: Culicidae). Journal of Vector Ecology 29, 340346.
  • Curtis CF (1992) Personal protection methods against vectors of disease. Review of Medical and Veterinary Entomology 80, 543553.
  • Curtis T & Williams DG (1994) Aromatic materials: I-Aroma chemicals. In: Introduction to Perfumery. Ellis Harwood, UK, pp. 140213.
  • Fradin MS (2001a) Protection from blood-feeding arthropods. In: Wilderness Medicine, 4th edn (ed PSAuerbach) Mosby, St. Louis, pp. 754768.
  • Fradin MS (2001b) Insect repellents. In: Comprehensive dermatologic drug therapy (ed SEWolverton) W.B. Saunders, Philadelphia, pp. 717734.
  • Fradin MS & Day JF (2002) Comparative efficacy of insect repellents against mosquito bites. The New England Journal of Medicine 347, 1318.
  • Golenda CF, Solberg VB, Burge R, Gambel JM & Wirtz RA (1999) Gender-related efficacy difference to an extended duration formulation of topical N, N-diethyl-m-toluamide (DEET). American Journal of Tropical Medicine and Hygiene 60, 654657.
  • Idem (1998) Mosquitoes and mosquito repellents: a clinician's guide. Annals of Internal Medicine 128, 931940.
  • Khan AA, Maibach HI & Skidmore DL (1975) Addition of vanillin to mosquito repellents to increase protection time. Mosquito News 35, 223225.
  • Limsuwan S, Rongsriyam Y, Kerdpibule V, Apiwathnasorn C, Chiang GL & Cheong WH (1987) Rearing techniques for mosquitoes. In: Practical Entomology. Malaria and filariasis, 1st edn (eds SucharitS & SupavejS) Museum and Reference Centre, Faculty of Tropical Medicine, Mahidol University, Bangkok.
  • Maibach HI, Khan AA & Akers WA (1974a) Use of insect repellents for maximum efficacy. Archives of Dermatology 109, 3235.
  • Maibach HI, Akers WA, Johnson HL, Khan AA & Skinner WA (1974b) Insects: topical insect repellents. Clinical Pharmacology and Therapeutics 16, 970973.
  • Palumbo JF & Lynn EV (1953) Dermatitis from celery. Journal of the American Pharmaceutical Association 42, 57.
  • Pauli G, Bessot JC, Dietmann-Molard A, Braun PA & Thierry R (1985) Celery sensitivity: clinical and immunological correlations with pollen allergy. Clinical Allergy 15, 273279.
  • Pitasawat B, Choochote W, Tuetun B et al. (2003) Repellency of aromatic turmeric Curcuma aromatica under laboratory and field conditions. Journal of Vector Ecology 28, 234240.
  • Qiu H, Jun HW & John WM (1998) Pharmacokinetic, formulation, and safety of insect repellent N,N-diethyl-3-methylbenzamide (DEET): a review. Journal of the American Mosquito Control Association 14, 1227.
  • Rafikali AM, Russel SR & Muraleednaran GN (2000) Bioactive compounds and 1,3-Di [(cis)-9-octadecenoyl]-2-[(cis,cis)-9-12-octadecadienoyl] glycerol from Apium graveolens L. seeds. Journal of Agricultural Food Chemical 48, 37853788.
  • Rafikali AM & Muraleednaran GN (2001) Mosquitocidal, nematocidal and antifungal compounds from Apium graveolens L. seeds. Journal of Agricultural Food Chemical 49, 142145.
  • Rattanarithikul R & Panthusiri P (1994). An illustrated key to the medically important mosquitoes of Thailand. US Army Medica; Companent, Southeast Asia Treaty Organization, Bangkok.
  • Robbins PJ & Cherniack MG (1986) Review of the biodistribution and toxicity of the insect repellent N,N-diethyl-m-toluamide (DEET). Journal of Toxicology and Environmental Health 18, 503525.
  • Sharma VP & Ansari MA (1994) Personal protection from mosquitoes (Diptera: Culicidae) by burning neem oil in kerosene. Journal of Medical Entomology 31, 95102.
  • Sharma SK, Dua VK & Sharma VP (1995) Field studies on the mosquito repellent action of neem oil. Southeast Asian Journal of Tropical Medicine and Public Health 26, 180182.
  • Tanaka K, Mizusawa K & Saugstad E (1979) A revision of the adult and larval mosquitoes of Japan (including the Ryukyu archipelago and the Ogasawara islands) and Korea (Dipetra: Culicidae). Contributions of the American Entomology Institue 16, pp. 987.
  • Tawatsin A, Wratten SD, Scott RR, Thavara U & Techadamrongsin Y (2001) Repellency of volatile oils from plants against three mosquito vectors. Journal of Vector Ecology 26, 7682.
  • Tuetun B, Choochote W, Rattanachanpichai E et al. (2004) Mosquito repellency of the seeds of celery (Apium graveolens L.). Annals of Tropical Medicine and Parasitology 98, 407417.
  • World Health Organization (1996) Report of the WHO informal consultation on the evaluation and testing of insecticides. CTD/WHOPES/IC/96.1, Control of Tropical Diseases Division. WHO, Geneva.
  • Yap HH, Jahangir K, Chong ASC et al. (1998) Field efficacy of a new repellent, KBR 3023, against Aedes albopictus (Skuse) and Culex quinquefasciatus (Say) in a tropical environment. Journal of Vector Ecology 23, 6268.