Background and Objectives
The sunn pest, Eurygaster integriceps Puton (Hemiptera: Scutelleridae), is a major pest of wheat and barley in Iran. Its geographic populations may vary in body size and shape due to environmental conditions and other factors. Traditional morphometrics relies on linear measurements, whereas geometric morphometrics analyzes shape using landmarks. Procrustes-transformed coordinates are widely used in geometric morphometric studies. However, researchers have rarely examined the value of residual shape data, obtained after accounting for size and allometric effects in sunn pest populations. This study compared traditional morphometrics, Procrustes data, and residual data to assess how well each distinguishes geographic populations of E. integriceps in Kurdistan Province, western Iran. It also assessed whether removing size-related effects improves population discrimination and whether the three approaches produce similar clustering patterns. Understanding geographic variation may provide a basis for future studies of population structure and improved pest management.
Materials and Methods
Adult sunn pests were collected from ten counties in Kurdistan Province, Iran: Sanandaj, Dehgolan, Ghorveh, Irankhah, Marivan, Divandareh, Tazehabad, Saghez, Bijar, and Kamyaran. After assessing specimen quality, 30 males and 30 females were randomly selected from each county. The final sample comprised 600 adults: 300 males and 300 females. Nine morphological characters were measured for traditional morphometric analysis. For geometric morphometrics, 18 landmarks were digitized on the dorsal surface of each specimen. Traditional measurements were standardized and analyzed using principal component analysis. We aligned landmark coordinates using Generalized Procrustes Analysis to remove differences in position, orientation, and scale. Centroid size was then calculated. Multivariate regression was used to assess allometry and obtain residual shape data after accounting for size-related effects. We conducted discriminant function analysis (DFA) separately for females and males using each of the three datasets. Classification accuracy was assessed through jackknifed cross-validation. Relationships among the ten geographic populations were also examined using average-linkage clustering (UPGMA) based on Mahalanobis distances.
Results
Discriminant function analysis (DFA) using traditional morphometric data from female and male insects revealed that, by cross-validation, 30.6% and 27.9% of the 10 sampled areas in Kurdistan Province were correctly discriminated for females and males, respectively. Similarly, DFA applying Procrustes data (geometric morphometrics) correctly discriminated 36.3% and 33.9% of the 10 sampled areas for females and males, respectively, based on cross-validation. In addition, DFA using residual data (geometric morphometrics) based on cross-validation correctly discriminated 42.9% and 39.6% of the 10 sampled areas for females and males, respectively. In the present study, the dendrograms generated from the residual and Procrustes data demonstrated similar clustering patterns for both females and males. Cluster analysis based on geometric morphometric data indicated that the dendrogram for females provided relatively better results than the dendrograms generated from traditional morphometric data for both sexes and from geometric morphometric data for males. In particular, Irankhah, Tazehabad, and Dehgolan were grouped with their geographically adjacent regions, Saghez, Divandareh, and Sanandaj, respectively. Furthermore, in DFA, residual data revealed higher discriminatory capacity than Procrustes data for differentiating the geographic populations of the sunn pest (E. integriceps) based on cross-validation.
Discussion
Geometric morphometrics distinguished the sampled geographic populations of E. integriceps more effectively than traditional morphometrics. The higher accuracy obtained with residual data suggests that accounting for centroid size and allometric effects can reveal shape differences that might otherwise be obscured. Residual data provided the highest cross-validated classification rates in both sexes, indicating that they can complement conventional Procrustes coordinates in population-level studies. The similar clustering patterns obtained from Procrustes and residual data also suggest that the main geographic relationships were broadly consistent across the two geometric approaches. Neighboring counties clustered together, matching their geographic proximity. However, classification rates remained below 50%, indicating that morphometric traits alone could not clearly distinguish all ten populations. Future studies could combine broader geographic sampling with independent evidence, such as molecular markers and environmental variables, to better understand the factors underlying population variation. Overall, residual geometric morphometric data provided the best discrimination, followed by Procrustes data and traditional morphometrics.