Table 1.
Adult acridid mortality (species, n = 31; family, n = 1; subfamily, n = 12) at different density levels (DL1 to DL5).
Fig 1.
Mortality (M%) in Acridids at different density levels (DL1 to DL5).
The upper and lower quartiles of each box represent the range of species mortality. M% for 31 species are represented by 31 dots in a box, where the line in the interquartile range represents the median value. Kruskal–Wallis test among DLs (F4, 154 = 146.133, p = 0.000, at 5%) reveals a significant difference in mortality across DLs. Dunn’s test analysis (0.05) reveals that M% differs significantly between DL1 and DL2 (p = 0.0) and DL2 and DL3 (P = 0.0), but not between DL3 and DL4 (p = 0.2894) and DL4 and DL5 (p = 0.63192).
Fig 2.
Distribution of acridid species based on mortality percentage (M%) at different density levels (DL1 to DL5).
Each bar represents the number of species (Y-axis) that displayed a specific M% (X-axis). Therefore, the sum of all bars for a DL represents the total number of acridid species (n = 31). Shifting bars from lower to higher mortality scales indicates that species tend more vulnerable to death as DL increases.
Fig 3.
Dose-response log-logistic regression model curves for 31 acridid species.
In this plot, the X-axis (independent variable) is referred to as ‘dose’ (species density in log scale = density concentration), and the Y-axis (dependent variable) is referred to as ‘response’ (mortality percentage of the species). Density concentrations (0.5, 1.0, 2.0, 4.0, and 8.0) for DL1 (10/20), DL2 (10/10), DL3 (10/5), DL4 (10/2.5) and DL5 (10/1.25) are thus plotted in log scale on the X-axis. The regression curves are developed based on species mortality at five density concentrations. The presented 31 dots for a single density concentration represent the 31 grasshopper species. The non-sigmoid mortality curves indicate that such species do not follow the rest pattern due to idiosyncratic (alike) mortalities across the DLs.
Fig 4.
A bi-plot of principal components (PC1, 57%; PC2, 31.6%) for 31 acridid species.
The PCA is classified according to the mortality percentages (M%) (response) acridids at different doses (DL1 to DL5). The cluster of DL4 and DL5 indicates a non-significant response, whereas the dispersed DLs (DL1, DL2, and DL3) indicate significant mortality changes.
Fig 5.
Hierarchical cluster relationship among 31 acridid species based on dose-response mortality values.
The species are grouped into five clusters (color sheds) based on their differential mortality exhibitions (response). Each frond of the dendrogram corresponds to objects (species) similar to each other, merged into branches, and fused at a higher height. The higher the height of the fusion, the less similar the species are, and the less similar the species are, the higher the fusion height.
Table 2.
Calculated effective concentrations (EC) for acridid species (n = 31).
Fig 6.
Pearson correlations between body mass index (BMI) of the species (n = 31) and mortality percentage (M%) at different density levels (DL1 to DL5).
The species are aligned in ascending order of BMI (from top to bottom) and corresponding M% at each DL. Correlation coefficients (r) are represented by circles ranging from -1.0 (red = negatively correlated) to +1.0 (green = positively correlated). The size of the circles corresponds to the value of r (the bigger the circle, the higher the r, and vice versa). Variation of one variable related to the variation of other is referred to as r2. An r2 of 0.5 indicates 25% of variations is correlated (0.5 squared = 0.25), and thus, for A gigantea, r for -0.99 (BMI vs. DL1) indicates that BMI of A gigantea is 99% negatively correlated with mortality at DL1. The value ‘zero’ (r = 0) suggests that there is no relationship between the variables (BMI and mortality) and ‘blank’ indicates that mortality at respective DL was zero (see Table 1).