Background

The Anopheles gambiae Giles complex is the most widely studied and the most important insect vector group. It is presently made up of eight sibling species: A. gambiae Giles (formerly molecular S form), A. coluzzii Coetzee and Wilkerson (formerly molecular M form), A. arabiensis Patton, the salt water breeders A. melas Theobald in West Africa and A. merus Dönitz in East Africa, A. quadriannulatus Theobald in southern Africa, A. amharicus Hunt Wilkerson and Coetzee and A. bwambae White in Uganda [1]. In Ghana, four species of the complex have been implicated as vectors of malaria; A. gambiae in the forest [2,3], A. melas at the coast, [4,5] and A. coluzzii and A. arabiensis in savanna areas [6,7]. A. melas and A. gambiae are vectors of lymphatic filariasis in coastal Ghana [8,9].

Identification and separation of the sibling species first became possible by using the morphology of polytene chromosomes of salivary gland cells of fourth instar larvae or the ovarian nurse cells of adult females [10]. Later, molecular techniques such as Polymerase Chain Reaction (PCR) were developed [11,12].

As one of the morphological tools, palp ratio (length of the fourth and fifth segment divided by length of the third) has been employed to separate A. melas from A. gambiae and A. arabiensis [13]. Bryan (1980) demonstrated that female Anopheles mosquitoes from The Gambia with palp ratios <0.81 and ≥0.81 could be identified as A. gambiae and A. melas respectively [14]. Using this cut-off point, the number of misidentified specimens was 3.8 % for A. melas and 5.8 % for A. gambiae. Later, Palsson et al. (1998) compared the palp ratio with PCR results and demonstrated for Anopheles from Guinea Bissau that a cut off point of 0.83 correctly identified 100 % of A. melas, but erroneously identified 4 % of A. gambiae as A. melas [15]. Akogbeto and Romano (1999) also reported the presence of A. gambiae and A. melas from coastal Benin and showed that a cut off ≤0.81 separates A. gambiae and >0.81 A. melas with an error of 3–6 % [16]. Palsson et al. (1998) therefore suggested that the PCR was the most optimal method available to separate A. gambiae and A. melas and that the palp ratio was not sufficiently reliable [15].

Within the framework of a study on lymphatic filariasis and intensity of transmission in the coastal area of Nzema East (Western Region), Ghana, we investigated whether the palp ratio method can be employed as a field tool to identify A. melas. Results of palp measurements were compared with those from Anopheles populations from the forest area of the Ashanti Region, Ghana.

Methods

A. gambiae s.l. mosquitoes were collected by Human Landing Catches (HLC) [17] from 18:00 hours to 02:00 hours in 6 villages along the sea coast near Essiama (Nzema East, Western Region) from September 2005 to January 2006. Volunteers (adult males) gave their verbal consent before participating in the collection of mosquitoes. Malaria prophylaxis was given, and treatment (at no cost to the volunteers) was arranged with the local hospitals but none became sick during the study period.

For comparison, A. gambiae females were received from two malaria transmission projects conducted in the forest areas of the Ashanti Region in two villages near Agona [2], and sites near Agogo and Konongo [18]. All mosquitoes were stored in cool boxes transferred to a field laboratory in Essiama, and dissected the following day for parity and infections with larval stages of Wuchereria bancrofti. The palps were removed, mounted in a drop of 1XPBS on a slide and segments 3 and 4 + 5 measured using a compound microscope. The legs and wings were stored in micro titre plates and transported to the laboratories of Kumasi Centre for Collaborative Research in Tropical Medicine (KCCR). They were stored at -20 °C for later DNA extraction and downstream PCR for species identification [11] and identified as A. gambiae s.s. and A. melas. The molecular forms M (now A. coluzzii) and S (A. gambiae) were not separated.

Data were entered using Microsoft Excel. The same programme was used in plotting scattergrams and frequency distributions. STATISTICA for Windows 1993 (StatSoft Inc., Tulsa, OK, USA) was used for the statistical analysis of the results. Correlation coefficients were determined by Pearson Product-Moment correlation and alpha values of less than 0.05 were considered significant.

Ethical approval

The study was part of a drug trial and treatment study [19] approved by the Ethical Committee of the School of Medical Sciences of the Kwame Nkrumah University of Science and Technology, Kumasi, Ghana. Study procedures were in accordance with the Helsinki Declaration of 1975 (as revised 1983 and 2000).

Results and discussion

Lengths of palp segments 3 and 4 + 5 of 1264 A. gambiae s.l. from the coastal area and 856 obtained from the forest zone [2,18] were measured. Lengths of segments 3 and 4 + 5 were significantly correlated in both samples (coastal area r = 0.75, p = 0.00, forest zone r = 0.60, p = 0.00, Pearson Product-Moment Correlation). The correlation coefficients of the two groups differed significantly (p = 0.00) indicating that the samples stemmed from different populations (Fig. 1). This was confirmed by the frequency distributions of the ratios of the lengths of palp segments 3 and 4 + 5 (Fig. 2).

Fig. 1
figure 1

Scattergram showing relations between length of segments 4 + 5 and segment 3 of 1400 A. melas from the study area and 856 A. gambiae s.s. from the control area

Full size image
Fig. 2
figure 2

Frequency distributions (average trends) of palp ratios (length segments 4 + 5/length segment 3) of 1264 A. melas from the study and 856 A. gambiae s.s. from the control area

Full size image

The majority of Anopheles mosquitoes from the coastal area were classified as A. melas. Outliers with palp ratios ≥ 0.81 were confirmed by PCR as A. melas. A. gambiae s.s., most likely the S form [5] of the fresh water breeders, is the only species of the A. gambiae complex known to occur in the Ashanti Region [2,9,20]. Therefore, using a palp ratio of 0.81 as the cut-off value, 14.9 % A. melas from the coastal area as confirmed by PCR would have been wrongly identified as A. gambiae. Similarly, using the same cut-off value and PCR, 10.2 % of specimens from the forest area would have been misidentified as A. melas (Table 1).

Table 1 Comparison of PCR and palp ratio (cut-off 0.81) for identification of putative A. melas from the Western Region and A. gambiae from Ashanti Region (controls) for species identification
Full size table

Conclusion

The palp ratio method proved to be a useful tool to identify populations of salt and freshwater A. melas and A. gambiae but not sufficiently reliable in identifying individual specimens. In the absence of PCR, especially in resource limited countries where students and scientists do not have access to molecular based techniques, this study recommends the use of the palp ratio to distinguish between A. gambiae and A. melas as vectors of malaria and lymphatic filariasis in the Ghana coastal area.