The present study was conducted at Burdwan (23° 16' N, 87° 54' E), West Bengal, India, during June-August 2006. Larvae of S. aegypti were obtained from a laboratory colony maintained in the Mosquito Research Unit, Department of Zoology, The University of Burdwan. The colony was kept free from exposure to pathogens, insecticides or repellents and maintained at 25–30°C. The larvae were fed on a powdered mixture of dog biscuits and dried yeast powder at a ratio of 3:1. The adult colony was provided with 10% sucrose solution and 10% multivitamin syrup, and was periodically blood-fed on restrained rats.
Preparation of aqueous extracts
Fresh, mature, green berries of S. villosum were randomly harvested during the study period from plants growing on the outskirts of Burdwan. All the berries were initially rinsed with distilled water and dried on a paper towel. The crude extracts were prepared by grinding the plant material in a mortar and pestle and passing the ground material through Whatman No 1 filter paper. Required concentrations of aqueous extracts were prepared by mixing the crude extract with a suitable amount of sterilized distilled water.
Preparation of plant extracts in different solvent systems
We harvested 25 g of fresh, mature berries, which were rinsed with distilled water and dried in a shed. The dried berries were put in a Soxhlet apparatus and the plant extracts were prepared using five solvents, namely petroleum ether, benzene, chloroform:methanol (1:1, v/v), acetone and absolute alcohol, applying one after another (extraction period 72 hour for each solvent and the temperature was < 40°C). The extracts were collected separately and the column of the Soxhlet apparatus was washed with 200 ml of water and 100 ml of a similar solvent as an eluent after each type of solvent extraction procedure. The eluted materials and each type of extract were concentrated in combination at 40°C to 100 ml of extract by evaporation in a rotary evaporator. Then each of the extracts was filtered, solvents were evaporated and the solid residues were weighed and then dissolved in a suitable amount of sterilized distilled water for the formulation of graded concentrations. The total yield of each extract from 25 g of berries was as follows: petroleum ether extract, 1.26 g; benzene extract, 2.38 g; chloroform:methanol (1:1, v/v) extract, 4.33 g; acetone extract, 3.00 g; and absolute alcohol extract 2.36 g.
The larvicidal bioassay followed the World Health Organization (WHO) standard protocols  with slight modifications. Each of the concentrations of aqueous berry extract (0.1–0.5%) was transferred into sterile glass Petri dishes (9 cm diameter/150 ml capacity). Ten third instar larval form of S. aegypti were separately introduced into different Petri dishes containing graded concentrations and the mortality were recorded after 24, 48 and 72 hours of the exposure period. The data of mortality in 48 and 72 hours were expressed by the addition of the mortality at 24 and 48 hours, respectively. Dead larvae were identified when they failed to move after probing with a needle in the siphon or cervical region. The experiments were replicated three times and conducted under laboratory conditions at 25–30°C and 80–90% relative humidity. Similar types of bioassay were conducted with different polar and non-polar solvent extracts (concentrations of 50, 25 and 15 ppm) of the green berries and with a chemical insecticide, Malathion, on third instar larval forms, and chloroform:methanol (1:1, v/v) extract on first and fourth instar larval forms. Larval toxicity was also tested according to similar methodologies using the bioactive substances (from chloroform:methanol extract) isolated from thin-layer chromatographic (TLC) plates.
The phytochemical analysis was carried out using the chloroform:methanol extract (as it exhibited highest mortality against S. aegypti larvae) of the green berries of S. villosum using the standard methods of Harbone  and Stahl . One or two drops of the chloroform:methanol extract were applied (using a capillary tube) to the bottom of each of the pre-coated and pre-heated (100°C for 30 minutes) glass plates (eight glass plates), which were prepared with silicagel G using Unoplan coating apparatus (Shadon, London). After 5 minutes of drying, each of the plates was placed in the separate glass chamber for TLC analysis, with different solvent systems as the mobile phase. After the movement of solvent at the top of the plates, each plate was removed from the glass chamber and separately air-dried. After 10 minutes each of plates was sprayed with a different spraying reagent for the identification of appropriate phytochemical. The phytochemicals included in the study were sapogenins, steroid, terpenoids, flavonoids, alkaloid, essential oils, phenolics and amino acids. A qualitative test was carried out to indicate the presence of saponins ; the remaining phytochemicals were determined using TLC analysis by the application of suitable solvents and spray reagents and, in each case, R
f values were recorded.
Ultraviolet-visual and infrared analysis of the active ingredient
The chloroform:methanol extract of the green berries of S. villosum was further chromatogrammed (30 plates) without the application of spraying reagents and each of the spots showed positive activity were separately scrapped according to their respective Rf values. Then each of the spots with their distinguishing Rf value was combined (from 30 plates) and undergoes further bioassay experiment to reveal the nature of active ingredient. As the spots exhibited positive response in Liberman Buchard reagent recorded highest larval mortality during further bioassay experiments, it undergoes spectral analysis by ultraviolet-visual (UV-Vis) and infrared (IR) spectroscopy. The UV-Vis analysis was carried out using a UV-1601 PC, SHIMADZU spectrophotometer with medium scan speed and sampling interval of 0.5 seconds. The IR spectroscopy analysis of the active spot was performed using KBr plates (JASCO FT-IR Model-420) with a scanning speed of 2 mm s-1.
All solvents and reagents used were of analytical grade and purchased from E. Merck, India. The TLC silica gel plates (0.25 mm thickness) were prepared and equilibrated with 2% (w/w) of water before use.
The percentage mortality observed (%M) was corrected using Abbott's formula  during the observation of the larvicidal potentiality of the plant extracts. A Student's t-test was performed to find the significance between the concentration of plant extract and mortality at different periods with different instars. Statistical analysis of the experimental data was performed using the computer software Statplus 2006 and MS EXCEL 2002 to find the LC50, regression equations (Y = mortality; X = concentrations) and regression coefficient values.