Morphometric geometric study of wing shape in Culex quinquefasciatus Say ( Diptera : Culicidae ) from Tamil Nadu , India

Author Details: K. MANIMEGALAI, is a Reader and the fields of specialization are: control of vector species of mosquitoes using plant extracts, mosquito taxonomy using wing shape analysis. M. ARUNACHALAM, is a Professor and his fields of specialization are: stream / river / wetlands ecology, fish ecology, fish habitat mapping using GIS, fish taxonomy and phylogenetics, morphometric geometry of shape analysis of the wings of aquatic insects and mosquitoes and shape analysis of cypriniform fishes. R. UDAYAKUMARI was a M.Phil., Scholar

from Mr. M. Murugan, Sri Paramakalyani Centre for Environmental Sciences of Manonmaniam Sundaranar University.We also acknowledge the help from Prof. James R. Rohlf for his reply to our questions while using his programmes.

I I I I INTRODUCTION NTRODUCTION NTRODUCTION NTRODUCTION NTRODUCTION
Landmark based geometric morphometrics have been increasingly used to explore systematic, developmental, ecological or pathological differences among individuals or populations, The approach of morphological landmarks is more appropriate as it represents advantages of additional information and the mode of variation is well understood (Bookstein 1989).Study of wing morphology in Lepidoptera and other insects using the geometric approach of shape variation provides an insight to taxonomy (Rohlf 1996).The genus Culex is a principal vector species in the peninsular India and the available keys are inadequate to study the morphological similarity of Culex species in this region and the present study is a pioneering attempt to distinguish the mosquito species based on wing morphometry.This paper forms a part of a major work on various aspects of different species of mosquitoes by the senior author.

M M M M MA A A
A ATERIALS TERIALS TERIALS TERIALS TERIALS AND AND AND

ETHODS ETHODS
Adult mosquitoes were collected from three localities in the Coimbatore city (the third largest city, 10 0 50'-11 0 02'N & 76 0 56'-77 0 01'E in the Indian state of Tamil Nadu with an area of 1287 km² with a population of 9,40,000 and the elevation is 442m above mean sea level).Wings were removed and photographed with consistent magnification and the digital images were used for the study.Landmarks (23) of the wings of 10 males and females of Culex quinquefasciatus were used for the analysis following the method of Rohlf & Slice (1990).The landmarks are the intersections of wing veins with the wing margin, intersections of cross veins with major veins, vein branch points , the apparent anterior attachment of the wing to the body , and the notch between the posterior wing margin and the alula (Figs.1-2).We digitized the photographs with the programme TPSDIG (Rohlf 1996a) and from the landmark coordinates we computed the consensus configuration using the programme TPSRELW (Rohlf 1997) which implements the algorithms described by Rohlf & Slice (1990) and Bookstein (1996).
For each wing the uniform component of shape variation (affine) is estimated and the first uniform component accounts for the stretching along the x-axis of the configuration whereas the 2 nd uniform component indicates dilations or compressions along the y-axis.The uniform components were estimated by the Linearized Procrustes method of Bookstein (1996b).All wings were scaled to unit centroid size before alignment by the GLS super imposition method.
Thin-plate spline of relative warp analysis was carried out using the coordinates of all aligned wings (Bookstein 1991;Rohlf 1993) and this will give the display and direction of shape differences among the two species (herein the male and female wings of Culex quinquefasciatus).The Thin-plate splines technique (Bookstein 1989) has the function of interpolation to the landmark coordinates of each wing against the reference configuration in order to make the landmarks homologous.The bending energy matrix  resulting from the thin-plate spline function fitted to the reference configuration will be modified as orthogonal axes which is called the principal warp which describe shape distortions of the reference configuration.The projection of the super imposed wings onto the principal warps produces the partial warp scores.These scores describe the deviations from the reference configuration.The relative warps are the principal components of the variation among wings (Bookstein 1989(Bookstein & 1991;;Rohlf 1993).The coverage configuration of landmarks are used as the reference which is aligned to its principal axis.The relative warps are computed with the scaling option = 0, which weights all landmarks equally.Otherwise all of the principal warps have the same weight and is considered to be more appropriate for systematic studies (Rohlf 1993;Rohlf et al. 1996).Weight matrix = 1 corresponds to the relative warp analysis which shows variation among wings in the principal warps with larger bending Deformation grids using Thin-plate splines were used to graphically represent the patterns of shape variation among the landmarks.We analyzed the warp matrix with multivariate analysis of variance (MANCOVA).

R R R R RESULTS ESULTS ESULTS ESULTS ESULTS
The positions of the landmark in the reference configuration are also indicated by the origin of vectors.The eigenvalue of the relative warp showed as 0.08482 and this is the square root of the variance per unit bending energy.The relative warp loadings showed values greater than zero (0.224, 0.145, 0.125, 0.1.2,0.77, 0.075, 0.023, 0.022 and 0.041) with the respective landmarks of 10,9,8,6,11,4,18,.The ordination of the non-affine shape of the first, second and third components of the Culex quinquefasciatus males and females are sharing the projections of partial warps towards the principal warps (Figs.6-7).Most of the variations between male and female are along the axis 1-2.
The first relative warp has more variations than the second because in axis one it is on the left side.In the second axis it is on the right side and in axis 3 showing extreme right and left.It is obvious  that the difference is between the base and tip of the wings.The affine components of males show variation of 93% and the females show 56% variation.The Thin-plate spline analysis shows that males show variation in the middle and distal ends while the female shows variation in the distal end of the wing.
It is also clear that the shape of the wings of male and female of Culex quinquefasciatus shows clear distinction based on landmark data with MANCOVA (Fig. 12).Two polygons are so distinct illustrating the clear difference between the wing shapes of males and females.

ISCUSSION ISCUSSION
The multivariate analysis of co-variance (MANCOVA) showed a clear separation of male and female mosquitoes of Culex quinquefasciatus based on wing shape.There is no overlap of any landmark data between the male and the female and the data of selected pairs of point are the distance measurements used as multidimensional space (Goodall 1991).The use of a parameter in this study following Rohlf (1993) is to identify appreciable changes in the value less than zero which will result in more weight to the small-scale features.In taxonomic studies = 0 gives an equal weighting in the analysis and in growth study to record the allometry where differences are more and hence = 1 is also used.As the present study reveals variations between male and female of Culex quinquefasciatus the study of using = 1 is tried to find out any interesting patterns at different scales.
The bending energy concept used in the study is the quantity of the physical properties of the thin metal sheets and used as Thin-plate spline and the functions used in organisms do not imply biological significance to the parameter.The use of spline (Rohlf 1993) does not imply that biological tissues behave like metal sheets but it is simply a convenient function that is able to express the differences in two configurations of landmarks as a continuous deformation.This Thin-plate spline is the assessment of shape deformations such as bio-orthogonal grids (Bookstein et al. 1985).
The Procrustes plots in the relative levels of variation at different levels are difficult to find out the displacement of landmarks.Hence, the method of relative warps has been used to identify the co-variation very effectively.From this study one can easily identify the movement of landmark either towards the base of the wings or towards the tip of the wings.

Figure 9 .
Figure 9. Plots of first two relative warps (based on Alpha = o) shown both as Thin -plate spline for positive and negative displacements along the first relative warp axes showing the net arrangement of land marks.(Alpha = 0)

Figure 11 .
Figure 11.The non -affine deformation for the comparison between Cluex quinquefasciatus male and female

Figure 12 .
Figure 12.Polygons formed on the projection of male and female wings on to the MANCOVA variables 1 and 2.