Correlation of Resistance Levels of Thrips flavus and Morphological Structures of Spring Soybean Varieties in Northeast China

Abstract

Thrips flavus (Thysanoptera: Thripidae) is a Eurasian pest that primarily attacks a variety of cash crops such as soybean. Currently, there is insufficient knowledge of thrips-resistance mechanisms in soybeans and a lack of effective thrips-resistant soybean varieties. The objective of this study was to identify the correlation between the pest thrips, T. flavus, resistance levels and morphological structures of soybean varieties. A total of 41 spring soybean varieties were planted in a field in Northeast China. Observations were made regarding the infestation intensity of T. flavus, the morphological structures (compound leaf shape, leaf length, leaf width, leaf surface humidity, trichome density, length, and color), leaf SPAD value, leaf nitrogen content, etc. Specifically, leaf trichome density (regardless of whether it was on the upper or lower surfaces of the upper, middle, or lower leaves), trichome color, and compound leaf shape all showed significant positive correlations with the amount of T. flavus. Additionally, principal component analysis (PCA) indicated that, during the peak flowering stage, leaf width, trichome length, trichome density, SPAD value, and nitrogen content were key factors for evaluating resistance; meanwhile, during the podding stage, leaf length, SPAD value, nitrogen content, and leaf surface humidity made the most significant contributions. Field resistance screening using the number of T. flavus per meter of double rows, the average number of T. flavus per plant, and hierarchical cluster analysis yielded consistent results. The soybean variety “podless-trichome” is a thrips-resistant variety (high resistance), and “Jinong 29” is a thrips-sensitive variety (high sensitivity). This study provides valuable insights into the occurrence of insect resistance to thrips in soybean varieties.

1. Introduction

Herbivorous thrips—including Neohydatothrips variabilis, Frankliniella tritici, F. fusca, Frankliniella occidentalis, and Thrips flavus—have become annual pests of soybeans in Brazil, North America, China, Egypt, and other regions [1,2,3,4]. The damage caused by thrips feeding and the spread of plant virus diseases are becoming increasingly severe [5,6]. Agricultural and chemical control are common preventive and management measures in integrated pest management (IPM) for thrips, a major pest of soybean. Insecticides continue to play an important role in controlling them [7,8]. Soybean thrips are susceptible to thiamethoxam, acephate, chlorfenapyr, and abamectin [2,9]. Lamiaceae and Rutaceae plant essential oils also showed effectiveness against these pests [10,11,12]. However, breeding insect-resistant varieties and discovering insect-resistant genes are the most effective and economical ways to control thrips damage [13]. The identification of germplasm resources that are resistant to thrips would be a valuable alternative strategy to the use of chemical insecticides [4]. The correlation of resistance levels among pests and morphological structures among plant varieties or genotypes is frequently the basis for breeding crops for resistance to pests [14]. Plant morphological and structural characteristics, such as leaf color, leaf thickness, pubescence (the presence of trichomes), and waxiness, influence pest behavior and population dynamics, consequently affecting host resistance against pests [15]. This is the first line of defense for plants against the feeding of pests. Plant chemicals also play a role in insect resistance. Plants can reduce their attractiveness to pests through changes in nutrient content, and nutrients can also participate in defense responses to enhance plant resistance to pests. Plant secondary metabolites have an impact on the behavior, feeding, digestion, and other aspects of pests [16]. Herbivorous thrips are generally considered to be a minor pest for soybean [3]. The insect resistance of soybean varieties to the pest thrips is still limited and unclear. Recently, Zhou et al. (2020) found that N. variabilis was less damaging in genotypes with low pubescence; all soybean genotypes exposed to soybean vein necrosis virus (SVNV)-infected N. variabilis exhibited typical SVNV symptoms and tested positive for SVNV, suggesting that none of the soybean genotypes tested were resistant to SVNV. They found that pubescence was a key component to N. variabilis feeding damage [15]. In a selective test of soybean infestation with uninfected and SVNV-infected N. variabilis, soybean varieties PI 229358 (Soden-Daizu) and PI 604464 (HC9515 MB) had the lowest number of wakame, and soybean varieties Williams 82 and Williamsfield Illini 3590 N had the highest number of adults [17]. The soybean genotypes showed exclusion-type resistance (antixenosis-type resistance) but not antibiosis-type resistance to both uninfected and SVNV-infected N. variabilis. It is therefore suggested that thrips resistance in soybean to N. variabilis may be related to biochemical responses or morphological traits, or a combination of both [17]. These results provide valuable reference information for the study of resistance to thrips in different soybean varieties. Further studies are also needed to characterize this mechanism of resistance and analyze the resistance of soybean genotypes to N. variabilis and SVNV [17].

T. flavus is a Eurasian pest that causes severe damage to a wide range of crops and is therefore of considerable economic importance [3,18,19,20]. This pest primarily attacks cash crops such as soybeans and maize, causing symptoms such as spotting, curling, chlorosis, and the wilting of leaves, which severely affect crop growth and yield [21]. It has also been suggested that yellow flower thrips may act as a vector for viruses such as tomato spotted wilt virus (TSWV), further exacerbating crop damage and posing a potential threat to agricultural production [18,22]. Due to its rapid reproductive rate and short generation cycle, it can quickly establish populations on crops and cause economic losses [23]. The effective control of T. flavus populations is therefore essential to ensuring crop yield and quality, maintaining the stability of agricultural ecosystems, and reducing economic losses to farmers [8]. Currently, there is insufficient knowledge of thrips-resistance mechanisms in soybean and a lack of effective thrips-resistant soybean varieties. Thus, we conducted a preliminary study on the relationship between the main morphological structures of spring soybean varieties and their resistance to T. flavus. The varieties’ resistance to thrips was evaluated using the thrips-abundance ratio, defined as the number of thrips per branch for each variety divided by the average number of thrips per branch across all varieties [24]. This study aimed to identify key leaf morphological and physiological traits associated with resistance to T. flavus in spring soybeans and to evaluate the resistance levels of 41 soybean varieties under field conditions in Northeast China. This study is of great importance in providing a scientific basis for the selection and breeding of thrips-resistant soybean varieties and the evaluation of varietal resistance to the insect; it also provides technical support for the integrated prevention and control of damage caused by T. flavus in soybean.

2. Materials and Methods

2.1. Test Soybean Varieties

The 41 spring soybean varieties (materials) selected for this study include Bei Dou 35, Bei Dou 3, He Feng 49, He Feng 53, Ji Mi Dou 1, Ji Nong 11, Ji Nong 18, Ji Nong 19, Ji Nong 20, Ji Nong 28, Ji Nong 29, Ji Nong 30, Ji Yu 20, Ji Yu 202, Ji Yu 203, Ji Yu 404, Ji Yu 47, Ji Yu 80, Ji Yu 82, Ji Yu 88, Ji Yu 90, Ji Yu 93, Ji Yu 95, Kang Xian Chong 12, Kang Xian Chong 6, Ken Dou 31, Ken Dou 33, Ken Feng 14, Ken Feng 15, Ken Feng 32, Long Huang 1, Jiu Qing Dou, Sui Nong 14, Sui Nong 28, Tong Nong 943, Yuan Yu 20, Za Jiao Dou 3, Chang Nong 25, Chang Nong 27, Chang Nong 34, and no pod trichome. The breeding research institutes and fertility periods associated with these 41 spring soybean varieties are provided in Table S1.

2.2. Design and Survey of the Experimental Plots

To investigate the occurrence of T. flavus on 41 spring soybean varieties under field conditions, this experiment was conducted from May to September 2022 at the Experimental Base of the Ministry of Agriculture and Rural Affairs (Jilin) Soybean Regional Technology Innovation Center (43°47′51.14″ N, 125°24′32.65″ E). The region has a continental monsoon climate characterized by dry and windy springs, hot and rainy summers, clear and pleasant autumns, and long, cold winters. The climate is characterized by distinct seasonal variations, synchronized rainfall and heat, and moderate dryness and humidity.

The experimental field was divided into plots, each measuring 3 m in length and 0.65 m in width. Each plot contained four rows of a single soybean variety, planted at a density of 30 seeds/m². A total of 41 soybean varieties were planted in the experiment, with three replicates for each variety, resulting in 123 plots in total. Planting was carried out on 6 May 2022 using manual point sowing (3 seeds per hole). Field management followed conventional agricultural practices; we undertook manual weeding, and no pesticides were applied throughout the growing season. The experiment was conducted without artificial irrigation, relying solely on natural rainfall.

To monitor the occurrence of T. flavus, regular surveys were conducted from the end of June until the end of the growing season. A total of 11 surveys were conducted from late June to mid-September 2022, covering the key growth stages of soybean, with intervals of 4–9 days. The specific survey dates were 21 June, 26 June, 30 June, 4 July, 9 July, 18 July, 25 July, 14 August, 20 August, 28 August, and 19 September 2022. The survey dates were scheduled based on key soybean growth stages (e.g., flowering and podding) and T. flavus infestation patterns, with adjustments made for weather conditions. This approach allowed us to capture biologically relevant trait–pest interactions. The survey method involved visual inspections. In each plot, a 1 m long bamboo stick was used to measure the middle 1 m double row of well-grown soybean plants with no missing seedlings. To ensure accuracy and reduce observer bias, each plot was surveyed independently by three trained observers, and the average value was used in the analysis. For each plot, the total number of soybean plants, the number of plants in the 1 m double row, and the number of adult and nymphal T. flavus were recorded.

2.3. Leaf Indicators Measurement

To investigate the morphological structures of the 41 spring soybean varieties, measurements were conducted during the peak flowering stage (14–24 July 2022) and the podding stage (18–20 August 2022). Five plants were randomly selected from each experimental plot for measurement. During both stages, one compound leaf was collected from the upper, middle, and lower layers of each plant, with three replications per layer.

Morphological structure indicators were measured according to the Description Specifications and Data Standard for Soybean (Glycine spp.) Germplasm Resources (by L.J. Qiu and R.Z. Chang) [25]. The measured indicators included leaf length, width, area, thickness, length-to-width ratio, trichome density and length, leaf color, compound leaf shape, plant height, stem diameter, plant architecture, number of effective branches, color of velvet on the upper and middle stems and pod husks, and flower color. Compound leaf shapes were categorized as lanceolate (1), ovate (2), elliptical (3), and round (4). Trichome colors were categorized as gray (1) and brown (2). Flower colors were categorized as white (1) and purple (2). Plant architectures were categorized as convergent (1) and semi-spreading (2). The colors of velvet on the upper and middle stem and pod husk were categorized as gray (1) and brown (2). Additionally, the values of soil–plant analytical development (SPAD), nitrogen content (N), leaf surface moisture, and leaf surface temperature were measured using a chlorophyll meter (TYS-4N, Beijing Zhongke Weihe Technology Development Co., Ltd., Beijing, China). Measurements were taken between 09:00 and 11:00 a.m. on the third fully expanded trifoliate leaf to minimize diurnal and developmental variability. Three readings per leaf were taken on five randomly selected plants per plot, and the values were averaged. The leaf morphological and physiological indicators and their measurement protocols are detailed in Table S2.

2.4. Principal Component Analysis (PCA) of Morphological Structure in 41 Spring Soybean Varieties

A higher trichome density may act as a physical barrier that hinders thrips movement or feeding, and trichomes may also secrete defensive secondary metabolites, forming a chemical defense. However, the positive correlation of a higher trichome density with the thrips number in our study suggests that the trichome’s morphology or genetic background may affect its defensive efficacy, or even attract certain populations, which requires further investigation. Wider leaves may provide a larger feeding area for thrips, increasing infestation risk, but they may also be associated with enhanced plant tolerance due to synergistic effects with other traits such as the nitrogen content and photosynthetic capacity. In addition, the SPAD value and nitrogen content may reflect both plant nutritional status and defense potential, while leaf humidity and temperature may influence thrips activity and population dynamics.

PCA was conducted to evaluate the morphological and physiological leaf traits of 41 spring soybean varieties at two key growth stages: peak flowering and podding. A total of 48 morphological indicators and 28 physiological indicators were analyzed across both stages.

At the peak flowering stage, 32 morphological and 16 physiological indicators were measured from new, upper, middle, and lower leaves. These indicators included leaf length, width, thickness, aspect ratio, surface area, trichome length, trichome density, SPAD value, nitrogen content, leaf surface humidity, and leaf surface temperature.

At the podding stage, data were collected from upper, middle, and lower leaves only, totaling 24 morphological and 12 physiological indicators. The same set of traits was evaluated to assess trait consistency across developmental stages.

PCA was used to reduce data dimensionality and identify the most influential variables contributing to morphological and physiological variation among the varieties.

2.5. Resistance Evaluation Methods

To evaluate the resistance of different soybean varieties to T. flavus, this study employed three resistance assessment methods for field resistance screening, based on the thrips resistance grading criteria established by previous researchers [24]:

(1) Thrips count in 1 m double rows: this method assesses resistance by comparing the proportion of thrips found in 1 m double rows of each soybean variety to the average total thrips count across all varieties.

(2) Average thrips per plant: the average number of thrips per plant for each soybean variety/the average total number of thrips per plant for all soybean varieties.

(3) Hierarchical cluster analysis: this method involves conducting a clustering analysis based on the survey data of T. flavus population dynamics for 41 spring soybean varieties. A total of 11 surveys were conducted to monitor the thrips population dynamics in the entire field, with each soybean variety having three replicates. The analysis was based on the average of the sum of thrips numbers from the three replicates for each survey.

According to the resistance evaluation grading criteria [24], the clustering results were classified into six levels: high resistance (HR), moderate resistance (MR), low resistance (LR), low susceptibility (LS), moderate susceptibility (MS), and high susceptibility (HS). The evaluation criteria for soybean thrips resistance are provided in Table S3.

2.6. Statistical Analysis of Data

Field trials were conducted to collect data on the population density of T. flavus and the morphological structure of each soybean variety. To accurately assess the infestation levels of T. flavus on different spring soybean varieties, the average values of the field population dynamic data were calculated. Subsequently, Pearson correlation analysis was performed using IBM SPSS Statistics 27 software (International Business Machines Corporation, Armonk, NY, USA) to examine the relationships between the morphological structures and the T. flavus populations of the 41 spring soybean varieties. All data used for the correlation analysis were mean values, allowing us to explore the relationships between the morphological characteristics of the varieties and the T. flavus populations. Additionally, cluster analysis was conducted using Origin 2021 software (OriginLab, Northampton, MA, USA) to reveal the similarities and differences among the 41 spring soybean varieties, providing a scientific basis for soybean pest-resistant breeding.

PCA was performed on the morphological structures of 41 spring soybean varieties. Data were first standardized, and the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s sphericity test were conducted using IBM SPSS Statistics 27. The KMO value was >0.500 (p < 0.05), indicating strong variable correlations and suitability for PCA. PCA was then carried out to extract principal components based on the loading matrix and scores. A biplot was generated using Python 3.10.11 (Python software, Python Software Foundation, Wilmington, DE, USA), with data preprocessing conducted via NumPy and Pandas and with visualization implemented using Matplotlib (v3.5.2) pyplot.

3. Results

3.1. Population Dynamics of T. flavus in 41 Spring Soybean Varieties

By conducting a survey of T. flavus population densities in the field, we found that soybean varieties with fewer T. flavus occurrences were the no pod trichome material and Bei Dou 3, while varieties with higher occurrences were Ji Nong 18, Ji Nong 29, and Ji Yu 202. Analysis of the T. flavus population dynamics across 41 soybean varieties revealed that T. flavus were present from the seedling stage to the beginning of the maturity stage, with the highest infestation levels occurring during the peak flowering period. Some varieties also showed high infestations during the pod initiation stage and podding stage. Specifically, the no pod trichome material reached its first peak on 18 July and its second peak on 28 August, with T. flavus counts of 23 and 18, respectively. The no pod trichome and Bei Dou 3 varieties consistently maintained the lowest T. flavus density across the season (Figure 1A). Bei Dou 3 reached its first peak on 18 July and its second peak on 20 August, with T. flavus counts of 45 and 34, respectively, after which the population gradually declined (Figure 1B). Among other soybean varieties, Bei Dou 35, Ji Yu 82, Ji Yu 404, Ken Dou 31, Ken Dou 33, Ken Feng 32, Long Huang 1, Chang Nong 27, and Chang Nong 34 reached their first peak on July 18, with T. flavus counts of 51, 46, 70, 44, 60, 41, 34, 79, and 71, respectively; they reached their second peak on 28 August, with counts of 28, 48, 33, 41, 34, and 61, respectively (Figures S1–S12). Ji Mi Dou 1, Ji Nong 11, Ji Nong 30, Ji Yu 20, Ji Yu 47, Ji Yu 80, Ji Yu 90, and Tong Nong 943 reached their first peak on 18 July, with T. flavus counts of 124, 84, 88, 63, 63, 108, 79, and 99, respectively; they reached their second peak on 28 August with counts of 34, 71, 59, and 46, respectively. He Feng 49, He Feng 53, and Ji Nong 19 reached their second peak on 25 July, with T. flavus counts of 53, 63, and 76, respectively. Ji Nong 28 reached three peaks on 4 July, 18 July, and 28 August, with T. flavus counts of 46, 47, and 64, respectively. Ji Yu 93, Kang Xian Chong 6, and Sui Nong 28 reached their peak on 25 July, with T. flavus counts of 103, 52, and 57, respectively. Yuan Yu 20 reached its peak on 28 August, with a T. flavus count of 57. Za Jiao Dou 3 reached three peaks on 30 June, 25 July, and 28 August, with T. flavus counts of 52, 50, and 54, respectively (Figures S13–S31). Ji Nong 20 and Ji Yu 95 reached their first peak on 18 July, with T. flavus counts of 94 and 80, respectively; Ji Nong 20 reached its second peak on 20 August, with a count of 73, and Ji Yu 95 reached its second peak on August 28, with a count of 70. Ji Yu 88 reached three peaks on 30 June, 18 July, and 14 August, with T. flavus counts of 63, 65, and 59, respectively. Ji Yu 203 reached three peaks on 4 July, 18 July, and 28 August, with T. flavus counts of 73, 107, and 46, respectively. Jiu Qing Dou reached three peaks on 4 July, 18 July, and 14 August, with T. flavus counts of 52, 73, and 71, respectively (Figures S32–S36). Ji Nong 18, Ji Nong 29, and Ji Yu 202 reached their first peak on July 18, with T. flavus counts of 127, 130, and 120, respectively; they reached their second peak on August 20, with counts of 65, 73, and 34, respectively, after which the population gradually declined (Figures S37–S39). The occurrence of multiple peaks or fluctuating thrips counts across different growth stages may be attributed to variations in growth stage, morphological structure, environmental factors, and resistance mechanisms.

3.2. Correlation Analysis of Morphological Structure and T. flavus Population Dynamics in 41 Spring Soybean Varieties

By conducting a correlation analysis between the morphological structure of 41 spring soybean varieties and T. flavus population numbers, we found that T. flavus numbers are primarily significantly correlated with soybean morphological indicators during the peak flowering stage but not significantly correlated with morphological indicators during the podding stage. During the peak flowering stage, the number of T. flavus is significantly negatively correlated with the aspect ratio of new leaves (p = 0.002 < 0.01, r = −0.463); the number of T. flavus is significantly negatively correlated with the length of upper leaves (p = 0.011 < 0.05, r = −0.392) and the aspect ratio of upper leaves (p = 0.009 < 0.01, r = −0.404) and significantly positively correlated with the density of velvet on the upper surface (p = 0.001 < 0.01, r = 0.514) and lower surface (p = 0.004 < 0.01, r = 0.441) of upper leaves. The number of T. flavus is significantly negatively correlated with the length of middle leaves (p = 0.007 < 0.01, r = −0.412), the aspect ratio of middle leaves (p = 0.004 < 0.01, r = −0.436), SPAD value (p = 0.022 < 0.05, r = −0.357), nitrogen content (p = 0.017 < 0.05, r = −0.372), and leaf surface humidity (p = 0.017 < 0.05, r = −0.370); it is significantly positively correlated with the density of velvet on the upper surface (p = 0.002 < 0.01, r = 0.473) and lower surface (p = 0.007 < 0.01, r = 0.416) of middle leaves. The number of T. flavus is significantly negatively correlated with the aspect ratio of lower leaves (p = 0.006 < 0.01, r = −0.418), SPAD value (p = 0.024 < 0.05, r = −0.353), nitrogen content (p = 0.021 < 0.05, r = −0.359), and leaf surface humidity (p = 0.021 < 0.05, r = −0.359); it is significantly positively correlated with the density of velvet on the upper surface (p = 0.011 < 0.05, r = 0.395) and lower surface (p = 0.034 < 0.05, r = 0.331) of lower leaves (Figure 2A–G).

Additionally, regardless of whether soybeans are in the peak flowering stage or the podding stage, the number of T. flavus is significantly positively correlated with the shape of compound leaves (p = 0.003 < 0.01, r = 0.451) and the color of velvet (p < 0.001, r = 0.574) (Figure 2H) and significantly positively correlated with the color of velvet on the upper and middle parts of the stem and pod skin (p < 0.001, r = 0.574) (Figure 2I). However, the number of T. flavus is not significantly correlated with leaf width, thickness, leaf area, trichome length, leaf surface temperature, plant height, stem diameter, effective branch number, or plant architecture, indicating that these morphological structures have little impact on the population dynamics of T. flavus. Specific detailed results are presented in Tables S4–S6.

3.3. PCA and Comprehensive Evaluation of Morphological Structure and T. flavus Population in 41 Spring Soybean Varieties

A PCA was conducted on the morphological structure and thrips population numbers of 41 spring soybean varieties, followed by a comprehensive evaluation of the selected principal components. PCA was employed to reduce the dimensionality of complex leaf trait datasets and identify key traits influencing T. flavus resistance, enabling a more comprehensive evaluation and ranking of soybean genotypes based on their morphological characteristics. The results showed that, during the peak flowering stage, the first seven leaf principal components and the first three leaf SPAD value principal components were chosen as comprehensive evaluation indicators for soybean varieties, with cumulative contribution rates of 86.065% and 82.898%, respectively. Among these, leaf width, trichome length, trichome density, leaf SPAD value, and leaf nitrogen content had the greatest positive impact on the principal components, indicating that these five morphological structures are crucial indicators for evaluating thrips resistance in soybean varieties during the peak flowering stage. The top three soybean varieties based on the comprehensive scores from the PCA of 32 leaf morphology indicators were Ji Yu 95, Chang Nong 27, and Tong Nong 943. For the 16 leaf SPAD value indicators, the top three varieties were Bei Dou 3, Bei Dou 35, and He Feng 53. During the podding stage, the first five leaf principal components and the first three leaf SPAD value principal components were selected as comprehensive evaluation indicators, with cumulative contribution rates of 85.277% and 95.203%, respectively. Leaf length, leaf SPAD value, leaf nitrogen content, and leaf surface humidity had the greatest positive impact on the principal components, indicating that these four morphological structures are crucial indicators for evaluating thrips resistance in soybean varieties during the podding stage. The top three soybean varieties based on the comprehensive scores from the PCA of 24 leaf morphology indicators were Wu Jiao Mao, Ji Yu 47, and Ji Yu 203. For the 12 leaf SPAD value indicators, the top three varieties were Ji Yu 20, Ji Nong 29, and Ji Yu 82 (Table S7).

3.3.1. PCA of Leaf Length, Width, Thickness, Length-to-Width Ratio, Leaf Area, Trichome Length, and Trichome Density

A PCA was conducted on the 32 leaf morphology indicators of 41 spring soybean varieties during the peak flowering stage, including leaf length, width, thickness, length-to-width ratio, leaf area, trichome length, and trichome density. First, the raw data were standardized, followed by a suitability test for sampling, which yielded a sampling adequacy value of 0.633 > 0.500. Bartlett’s test of sphericity was also performed, resulting in a non-parametric test statistic value of 1875.075 with p < 0.05, further indicating strong correlations among the indicators, making PCA appropriate. Based on the criterion of eigenvalues greater than 1, the first seven principal components were selected. The results are shown in Figure 3, with a cumulative contribution rate of 86.065%, capturing the majority of the information from the eight morphological structures. Therefore, these seven principal components can be used as comprehensive evaluation indicators for soybean varieties.

From Table S7, it is evident that the first principal component has the highest contribution rate of 30.641%, with an eigenvalue of 9.805, indicating its dominant role in the analysis and evaluation. In the first principal component, positive influences come from the width of new leaves, upper, middle, and lower leaves; the leaf area of upper, middle, and lower leaves; the trichome length of new, upper, and middle leaves; and trichome density on both the upper and lower surfaces of new, upper, middle, and lower leaves. This suggests that leaf width, trichome length, and trichome density are crucial indicators for evaluating thrips resistance in soybean varieties (Figure 3).

The second principal component has a contribution rate of 18.922% and an eigenvalue of 6.077. Positive influences include the width, length, thickness, and leaf area of new, upper, middle, and lower leaves, as well as trichome density on both the upper and lower surfaces of new leaves.

The third principal component has a contribution rate of 11.506% and an eigenvalue of 3.682. Positive influences include the width, length, thickness, length-to-width ratio, and leaf area of new, upper, middle, and lower leaves, as well as the trichome length and trichome density on the upper and lower surfaces of upper, middle, and lower leaves. Negative influences are exerted by trichome density on both the upper and lower surfaces of new leaves.

The fourth principal component has a contribution rate of 8.024% and an eigenvalue of 2.568. Positive influences include the width of upper leaves; the length of new, upper, middle, and lower leaves; the length-to-width ratio of new, upper, middle, and lower leaves; the leaf area of upper, middle, and lower leaves; and trichome density on both the upper and lower surfaces of new, upper, middle, and lower leaves.

The fifth principal component has a contribution rate of 7.459% and an eigenvalue of 2.387. Positive influences include the width of lower leaves; length of middle and lower leaves; thickness of new, upper, middle, and lower leaves; length-to-width ratio of new, upper, middle, and lower leaves; leaf area of middle and lower leaves; and trichome density on both the upper and lower surfaces of new, upper, middle, and lower leaves.

The sixth principal component has a contribution rate of 5.151% and an eigenvalue of 1.648. Positive influences include the width of new and lower leaves; length of new, middle, and lower leaves; thickness of middle leaves; length-to-width ratio of new, upper, middle, and lower leaves; leaf area of new and lower leaves; and trichome density on both the upper and lower surfaces of upper, middle, and lower leaves.

The seventh principal component has a contribution rate of 4.291% and an eigenvalue of 1.373. Positive influences include the width of new and upper leaves; length of upper leaves; thickness of new, upper, and middle leaves; length-to-width ratio of middle and lower leaves; leaf area of new and upper leaves; trichome length of new leaves; and trichome density on both the upper and lower surfaces of upper, middle, and lower leaves.

3.3.2. PCA of Leaf SPAD Value, Nitrogen Content, Leaf Surface Humidity, and Leaf Surface Temperature

A PCA was conducted on the 16 leaf indicators of 41 spring soybean varieties during the peak flowering stage, including the leaf SPAD value, nitrogen content, leaf surface humidity, and leaf surface temperature. First, the raw data were standardized, followed by a suitability test for sampling, which yielded a sampling adequacy value of 0.813 (>0.500). Bartlett’s test of sphericity was also performed, resulting in a non-parametric test statistic value of 2343.124 (p < 0.05), indicating strong correlations among the indicators, meaning that PCA is appropriate. Based on the criterion of eigenvalues greater than 1, the first three principal components were selected, with a cumulative contribution rate of 82.898%. The results are shown in Table S8 and Figure 4.

The first principal component has a contribution rate of 56.769%, with an eigenvalue of 9.083. High loadings are observed for the SPAD values and nitrogen content of new leaves and upper, middle, and lower leaves; these two indicators have a positive impact on the first principal component.

The second principal component has a contribution rate of 16.392%, with an eigenvalue of 2.623. Positive influences include the SPAD values and nitrogen content of new and upper leaves, as well as the leaf surface humidity and temperature of new leaves and upper, middle, and lower leaves. Negative influences are exerted by the SPAD values and nitrogen content of middle and lower leaves, as well as the leaf surface humidity of upper, middle, and lower leaves.

The third principal component has a contribution rate of 9.738%, with an eigenvalue of 1.588. Positive influences include the SPAD values and nitrogen content of middle and lower leaves, as well as the leaf surface humidity of middle and lower leaves and the leaf surface temperature of new, upper, middle, and lower leaves.

3.3.3. Comprehensive Evaluation of 32 Leaf Indicators for 41 Spring Soybean Varieties During Peak Flowering Stage

Based on the eigenvalues of each principal component in Table S9, the principal component scores for the 41 spring soybean varieties were calculated. Using the contribution rate of each principal component as a weight, a comprehensive evaluation model for different soybean varieties was constructed:

$$\mathrm{y}=0.306\times {\mathrm{y}}_1+0.190\times {\mathrm{y}}_2+0.115\times {\mathrm{y}}_3+0.080\times {\mathrm{y}}_4+0.075\times {\mathrm{y}}_5+0.052\times {\mathrm{y}}_6+0.043\times {\mathrm{y}}_7$$

According to Table S9, the soybean varieties with a composite score greater than 1 are Ji Yu 95, Chang Nong 27, Tong Nong 943, Ji Yu 82, Ji Nong 30, Ji Yu 202, and Ji Yu 80. The soybean varieties with a composite score less than −1 are Bei Dou 3, Yuan Yu 20, Ji Yu 404, Ken Dou 31, Ji Yu 20, Ken Feng15, Ji Yu 203, Hei Feng 53, no pod trichome, Sui Nong 28, and Long Huang 1.

3.3.4. Comprehensive Evaluation of 16 Leaf Indicators for 41 Spring Soybean Varieties During Peak Flowering Stage

Based on the eigenvalues of each principal component in Table S10, the principal component scores for the 41 spring soybean varieties were calculated. Using the contribution rate of each principal component as a weight, a comprehensive evaluation model for different soybean varieties was constructed:

$$\mathrm{y}=0.5677\times {\mathrm{y}}_1+0.1640\times {\mathrm{y}}_2+0.0974\times {\mathrm{y}}_3$$

According to Table S10, the soybean varieties with a composite score greater than 1 are Bei Dou 3, Bei Dou 35, He Feng 53, Sui Nong 14, Ji Yu 82, Ken Dou 33, Ji Yu 404, Ji Yu 20, Ji Nong 18, Ji Nong 19, Ken Feng 14, and Tong Nong 943. Conversely, the soybean varieties with a composite score less than −1 are Ji Yu 203, Kang Xian Chong 12, no pod trichome, Ji Mi Dou 1, Chang Nong 25, Ji Nong 30, and Ji Yu 80.

3.3.5. PCA of Morphological Structure Indicators for 41 Spring Soybean Varieties During the Podding Stage

(1) PCA of Leaf Length, Width, Thickness, Length-to-Width Ratio, Leaf Area, Trichome Length, and Trichome Density

A PCA was conducted on the 24 leaf indicators of 41 spring soybean varieties during the podding stage, including leaf length, width, thickness, length-to-width ratio, leaf area, trichome length, and trichome density. First, the raw data were standardized; this was followed by a suitability test for sampling, which yielded a sampling adequacy value of 0.681 (>0.500). Bartlett’s test of sphericity was also performed, resulting in a non-parametric test statistic value of 1653.621 (p < 0.05), further indicating strong correlations among the indicators, making PCA appropriate.

Based on the criterion of eigenvalues greater than 1, the first five principal components were selected. The results are shown in Table S11, with a cumulative contribution rate of 85.277%. Therefore, these five principal components can be used as comprehensive evaluation indicators for soybean varieties.

According to Table S11 and Figure 5, the contribution rate of principal component 1 is the highest at 36.890%, with an eigenvalue of 8.854. The upper leaf length, middle leaf length, and lower leaf length have high loadings and produce a positive impact on this principal component.

Principal component 2 has a contribution rate of 27.808%, with an eigenvalue of 6.674. Positive influences include lower leaf length, upper leaf width, middle leaf width, lower leaf width, upper leaf thickness, lower leaf thickness, upper leaf area, middle leaf area, lower leaf area, upper trichome length, upper leaf upper surface trichome density, middle leaf upper surface trichome density, lower leaf upper surface trichome density, upper leaf lower surface trichome density, middle leaf lower surface trichome density, and lower leaf lower surface trichome density. Negative influences are as follows: upper leaf length, middle leaf length, middle leaf thickness, upper leaf length-to-width ratio, middle leaf length-to-width ratio, lower leaf length-to-width ratio, middle trichome length, and lower trichome length.

Principal component 3 has a contribution rate of 9.425%, with an eigenvalue of 2.262. Positive influences include the following: middle leaf width, upper leaf thickness, middle leaf thickness, upper trichome length, middle trichome length, lower trichome length, upper leaf lower surface trichome density, and lower leaf lower surface trichome density.

Principal component 4 has a contribution rate of 6.598%, with an eigenvalue of 1.583. Positive influences include upper leaf length, middle leaf length, lower leaf length, lower leaf width, upper leaf thickness, middle leaf thickness, lower leaf thickness, middle leaf area, lower leaf area, upper leaf length-to-width ratio, middle leaf length-to-width ratio, lower leaf length-to-width ratio, middle trichome length, lower trichome length, upper leaf upper surface trichome density, middle leaf upper surface trichome density, lower leaf upper surface trichome density, upper leaf lower surface trichome density, middle leaf lower surface trichome density, and lower leaf lower surface trichome density.

Principal component 5 has a contribution rate of 4.555%, with an eigenvalue of 1.093. Positive influences include the following: upper leaf length, lower leaf length, upper leaf width, lower leaf width, upper leaf thickness, lower leaf thickness, upper leaf area, lower leaf area, upper leaf length-to-width ratio, middle leaf length-to-width ratio, lower leaf length-to-width ratio, upper trichome length, middle trichome length, lower trichome length, middle leaf upper surface trichome density, lower leaf upper surface trichome density, middle leaf lower surface trichome density, and lower leaf lower surface trichome density.

(2) PCA of Leaf SPAD Value, Nitrogen Content, Leaf Surface Humidity, and Leaf Surface Temperature

A PCA was conducted on the 12 leaf indicators of 41 spring soybean varieties during the peak flowering stage, including leaf SPAD value, nitrogen content, leaf surface humidity, and leaf surface temperature. First, the raw data were standardized; this was followed by a suitability test for sampling, which yielded a sampling adequacy value of 0.811 (>0.500). Bartlett’s test of sphericity was also performed, resulting in a non-parametric test statistic value of 2712.297 (p < 0.05), indicating strong correlations among the indicators, making PCA appropriate. Based on the criterion of eigenvalues greater than 1, the first three principal components were selected, with a cumulative contribution rate of 95.203%, as shown in Table S12.

Principal component 1 has the highest contribution rate of 59.830%, with an eigenvalue of 7.180. The upper leaf SPAD value, middle leaf SPAD value, lower leaf SPAD value, upper leaves nitrogen content, middle leaves nitrogen content, lower leaves nitrogen content, upper leaf surface humidity, middle leaf surface humidity, and lower leaf surface humidity have high loadings and produce a positive impact on this principal component. Principal component 2 has a contribution rate of 22.927%, with an eigenvalue of 2.751. Positive influences include lower leaf SPAD value, lower leaf nitrogen content, lower leaf surface humidity, upper leaf surface temperature, middle leaf surface temperature, and lower leaf surface temperature. Principal component 3 has a contribution rate of 12.446%, with an eigenvalue of 1.494. Positive influences include upper leaf SPAD value, upper leaf nitrogen content, upper leaf surface humidity, upper leaf surface temperature, middle leaf surface temperature, and lower leaf surface temperature.

(3) Comprehensive Evaluation of 24 Leaf Indicators for 41 Spring Soybean Varieties During the Full Pod Stage

Based on the eigenvalues of each principal component listed in Table S13, the principal component factor scores for the 41 spring soybean varieties were calculated. Using the contribution rate of each principal component as a weight, a comprehensive evaluation model was constructed for different soybean varieties:

$$\mathrm{y}=0.3689\times {\mathrm{y}}_1+0.2781\times {\mathrm{y}}_2+0.0943\times {\mathrm{y}}_3+0.0660\times {\mathrm{y}}_4+0.0456\times {\mathrm{y}}_5$$

According to Table S13, the soybean varieties with a composite score greater than 1 are no pod trichome, Ji Yu 47, Ji Yu 203, Chang Nong 27, Ji Yu 95, Ji Yu 82, Jiu Ying Dou, Ji Nong 20, and Za Jiao Dou 3. The soybean varieties with a composite score less than −1 are Ji Yu 90, Ken Feng 14, Bei Dou 3, Chang Nong 25, Ji Nong 29, Sui Nong 14, Hei Feng 49, Ji Yu 202, Ken Dou 31, and Ji Mi Dou 1.

(4) Comprehensive Evaluation of 12 Leaf Indicators for 41 Spring Soybean Varieties During the Full Pod Stage

Based on the eigenvalues of each principal component listed in Table S14, the principal component factor scores for the 41 spring soybean varieties were calculated. Using the contribution rate of each principal component as a weight, a comprehensive evaluation model was constructed for different soybean varieties:

$$\mathrm{y}=0.5983\times {\mathrm{y}}_1+0.2293\times {\mathrm{y}}_2+0.1245\times {\mathrm{y}}_3$$

According to Table S14, the soybean varieties with a composite score greater than 1 are Ji Yu 20, Ji Nong 29, Ji Yu 82, Ji Yu 47, Ken Feng 14, Ji Yu 404, Sui Nong 28, and Ken Dou 31. The soybean varieties with a composite score less than −1 are Ken Feng 15, Long Huang 1, He Feng 49, Ji Yu 88, Jiu Qing Dou, Ji Yu 93, Ken Feng 32, Jinong 18, and no pod trichome.

3.4. Field Resistance Screening of 41 Spring Soybean Varieties Against T. flavus

To determine the field resistance of 41 spring soybean varieties against T. flavus, this study used three methods for evaluation: the number of T. flavus per meter of double rows, the average number of T. flavus on each plant, and hierarchical cluster analysis. These methods were employed to enhance the stability of the results. The comprehensive evaluation using these three resistance assessment methods revealed that, while there were some differences in the results, the overall trends were consistent. The average number of T. flavus on each plant provided a more comprehensive evaluation of soybean resistance against T. flavus and yielded more reliable results. The population dynamics of T. flavus on different resistant soybean varieties are shown in Figure 6A–F. Based on the average number of T. flavus on each plant, the field performance of highly resistant soybean varieties is characterized by the absence of pod trichomes. Moderately resistant varieties include Bei Dou 3, and less resistant varieties include Bei Dou 35, Ji Yu 404, Ji Yu 82, Kang Xian Chong 12, Kan Dou 31, Kan Dou 33, Ken Feng 15, Ken Feng 32, Long Huang 1, Chang Nong 25, Chang Nong 27, and Chang Nong 34. Slightly susceptible varieties include Hei Feng 49, Hei Feng 53, Ji Mi Dou 1, Ji Nong 11, Ji Nong 19, Ji Nong 28, Ji Nong 30, Ji Yu 20, Ji Yu 47, Ji Yu 80, Ji Yu 90, Ji Yu 93, Kang Xian Chong 6, Ken Feng 14, Sui Nong 14, Sui Nong 28, Tong Nong 943, Yuan Yu 20, and Za Jiao Dou 3. Moderately susceptible varieties include Ji Nong 20, Ji Yu 203, Ji Yu 88, Ji Yu 95, and Jiu Qing Dou. Highly susceptible varieties include Ji Nong 18, Ji Nong 29, and Ji Yu 202. Specifically:

(1) The resistance results based on the number of T. flavus per meter of double rows are shown in

Table 1. Investigation of resistance to T. flavus in 41 soybean varieties.

Resistance Levels	Soybean Variety	Thrips Amount Ratio	Soybean Variety	Thrips Amount Ratio
High Resistance (HR)	No pod trichome	0.2
Medium Resistance (MR)	Bei Dou 3	0.48	Long Huang 1	0.57
	Bei Dou 35	0.54	Ken Dou 31	0.59
Low Resistance (LR)	He Feng 49	0.63	Ji Yu 47	0.84
	Sui Nong 28	0.63	Yuan Yu 20	0.84
	Ken Feng 15	0.64	Ji Yu 90	0.91
	Ji Yu 82	0.65	Ji Yu 202	0.93
	Ken Dou 33	0.65	Ji Nong 30	0.95
	Ken Feng 32	0.67	Ji Yu 203	0.96
	He Feng 53	0.71	Jiu Qing Dou	0.96
	Sui Nong 14	0.71	Ji Yu 88	0.98
	Ken Feng 14	0.73	Ji Nong 20	1
	Ji Nong 28	0.74	Ji Yu 93	1.02
	Ji Nong 11	0.75	Chang Nong 34	1.02
	Ji Yu 20	0.75	Ji Yu 95	1.03
	Ji Yu 404	0.75	Ji Yu 80	1.04
	Kang Xian Chong 12	0.76	Ji Mi Dou 1	1.07
	Chang Nong 25	0.77	Ji Nong 18	1.17
	Tong Nong 943	0.78	Ji Nong 29	1.34
	Kang Xian Chong 6	0.83
Low Sensitivity (LS)	Ken Dou 31	0.67	Yuan Yu 20	1.04
	Bei Dou 35	0.71	Ji Yu 90	1.06
	Ken Feng 32	0.81	Ji Mi Dou 1	1.08
	Chang Nong 25	0.85	He Feng 53	1.11
	Kang Xian Chong 12	0.88	Ji Yu 93	1.11
	Chang Nong 27	0.89	Kang Xian Chong 6	1.11
	Ji Yu 404	0.9	Ji Nong 30	1.12
	Chang Nong 34	0.9	Ji Nong 11	1.18
	Sui Nong 14	0.92	Ji Yu 80	1.19
	He Feng 49	0.93	Zajiao Dou 3	1.19
	Ji Nong 28	0.96	Ji Nong 20	1.24
	Ji Yu 20	0.96	Ji Yu 88	1.3
	Ken Feng 14	0.96	Ji Yu 95	1.3
	Ji Nong 19	0.99	Ji Yu 203	1.44
	Sui Nong 28	0.99	Jiu Qing Dou	1.45
	Tong Nong 943	1	Ji Yu 202	1.51
	Ji Yu 47	1.03
Medium-Sensitive (MS)	Ji Nong 20	1.24	Ji Yu 95	1.3
	Ji Yu 88	1.3
Highly Sensitive (HS)	Ji Nong 18	1.53	Ji Nong 29	1.53

. One variety showed high resistance in the field, no pod trichome, accounting for 2.43%. Four varieties showed moderate resistance: Bei Dou 35, Bei Dou 3, Kan Dou 31, and Long Huang 1, accounting for 9.75%. Moreover, 22 varieties showed low resistance: Hei Feng 49, Hei Feng 53, Ji Nong 11, Ji Nong 19, Ji Nong 28, Ji Yu 20, Ji Yu 404, Ji Yu 47, Ji Yu 82, Kang Xian Chong 12, Kang Xian Chong 6, Kan Dou 33, Ken Feng 14, Ken Feng 15, Ken Feng 32, Sui Nong 14, Sui Nong 28, Tong Nong 943, Yuan Yu 20, Za Jiao Dou 3, Chang Nong 25, and Chang Nong 27, accounting for 53.65%. Then, 13 varieties showed slight susceptibility: Ji Mi Dou 1, Ji Nong 18, Ji Nong 20, Ji Nong 30, Ji Yu 202, Ji Yu 203, Ji Yu 80, Ji Yu 88, Ji Yu 90, Ji Yu 93, Ji Yu 95, Jiu Qing Dou, and Chang Nong 34, accounting for 31.70%. One further variety showed moderate susceptibility, Ji Nong 29, accounting for 2.43%.

(2) The resistance results based on the average number of T. flavus on each plant are shown in Table 1. One variety showed high resistance in the field, no pod trichome, accounting for 2.43%. One variety showed moderate resistance, Bei Dou 3, accounting for 2.43%. A total of 12 varieties showed low resistance: Bei Dou 35, Ji Yu 404, Ji Yu 82, Kang Xian Chong 12, Kan Dou 31, Kan Dou 33, Ken Feng 15, Ken Feng 32, Longhuang 1, Chang Nong 25, Chang Nong 27, and Chang Nong 34, accounting for 29.26%. A total of 19 varieties showed slight susceptibility: Hei Feng 49, Hei Feng 53, Ji Mi Dou 1, Ji Nong 11, Ji Nong 19, Ji Nong 28, Ji Nong 30, Ji Yu 20, Ji Yu 47, Ji Yu 80, Ji Yu 90, Ji Yu 93, Kang Xian Chong 6, Ken Feng 14, Sui Nong 14, Sui Nong 28, Tong Nong 943, Yuan Yu 20, and Za Jiao Dou 3, accounting for 46.34%. Five varieties showed moderate susceptibility: Ji Nong 20, Ji Yu 203, Ji Yu 88, Ji Yu 95, and Jiu Qing Dou, accounting for 12.19%. Three varieties showed high susceptibility: Ji Nong 18, Ji Nong 29, and Ji Yu 202.

(3) The resistance classification based on hierarchical cluster analysis is shown in Figure 7. One variety showed high resistance in the field, no pod trichome, accounting for 2.4%. Eight varieties showed moderate resistance, accounting for 19.5%: Ji Nong 11, Ji Nong 19, Tong Nong 943, Kang Xian Chong 6, Chang Nong 27, Ji Yu 203, Ji Yu 80, and Ji Yu 93. Seven varieties showed low resistance, accounting for 17.1%: Ji Nong 20, Ji Yu 90, Ji Nong 30, Jiu Qing Dou, Ji Yu 95, Chang Nong 34, and Ji Yu 88. In total, 21 varieties showed slight susceptibility, accounting for 51.2%: Chang Nong 25, Hei Feng 53, Za Jiao Dou 3, Ji Yu 47, Yuan Yu 20, Ji Nong 28, Ji Yu 404, Ji Yu 20, Sui Nong 28, Ken Feng 14, Sui Nong 14, Kan Dou 33, Kang Xian Chong 12, Ken Feng 32, Ji Yu 82, Hei Feng 49, Ken Feng 15, Long Huang 1, Bei Dou 3, Kan Dou 31, and Bei Dou 35. Two varieties showed moderate susceptibility, accounting for 4.9%, Ji Mi Dou 1 and Ji Yu 202. Two further varieties showed high susceptibility, accounting for 4.9%, Ji Nong 18 and Ji Nong 29.

4. Discussion

In this study, we conducted a field survey on the population size of T. flavus on 41 spring soybean varieties at the flowering and podding stages, determined the main morphological structure indexes of the 41 spring soybean varieties, and performed correlation analysis and PCA on the population size and morphological structure of T. flavus. The correlation analysis and PCA were conducted to evaluate the thrips resistance of the 41 spring soybean varieties based on the correlation results and the composite scores of the principal components.

In the investigation, T. flavus occurred from the seedling stage to the onset of maturity in soybean, and the maximum occurrence of thrips was detected at the peak flowering stage, which indicated that the peak population size occurred at the flowering stage; the damage thrips cause to soybeans during this period might also be at its maximum, which is consistent with the results of Gao et al. (2019) [26]. This coincides with an increase in the availability of high-protein food sources such as soybean pollen [4]. This period contributes to significant population growth due to the abundance of food resources available. Additionally, the arrival of the rainy season affects thrips population dynamics: the more rainfall there is, the smaller the population tends to be. These aspects will be further explored in our future research.

Plant structural traits represent an important component of plant defense against herbivory, often acting as the first line of resistance through adaptations such as waxy cuticles, pubescence, and leaf toughening [27,28]. These morphological features can limit insect feeding and movement, thereby reducing damage. Epidermal trichomes, in particular, are among the major protective tissues in plants and play a significant role in defending against both biotic and abiotic stresses [29]. The influence of soybean leaf trichome density on feeding preference and development has been proven for several pests, including aphids, leafhoppers, beetles, and whiteflies [30,31,32,33]. In this study, among the morphological parameters analyzed, only the density of velvet hairs played a role in providing resistance to T. flavus. With similar results, chili (Capsicum annuum) varieties with high resistance to Scirtothrips dorsalis and Polyphagotarsonemus latus had the highest velvet density [34]. Rahman et al. (2022) observed that physical characteristics such as trichome density on the leaf lamina, midrib, and petiole showed negative correlations with the incidence of thrips in groundnut [35]. Usman et al. (2020) observed a negative relationship between S. dorsalis density and trichome density and type in five commercially cultivated tomato genotypes (“Riogrande,” “Riogrande H,” “Bombino,” “Roma V’F,” and “Roma”) [36]. Zhao et al. (2008) investigated the relationship between soybean trichome density and resistance to soybean pod borer and found that soybean cultivars with lower trichome density were more resistant to Leguminivora glycinivorella [37]. Soybean insect resistance was significantly negatively correlated with trichome color, trichome density, and trichome length under natural infestation conditions [38]. Xu et al. (2005) and Dibbad et al. (2022) found a significant correlation between leaf trichome characteristics and tobacco fly resistance, with trichomeless leaves being a resistant trait [39,40]. In this study, we found that the lower the density of leaf trichomes, the lower the number of thrips and the higher the resistance, which is consistent with the above findings. This may be due to the small body size of thrips and the large number of trichomes providing a strong barrier effect, which is favorable for the safe feeding and reproduction of thrips. There are also studies that do not agree with the results of the present study. Lee et al. (1986) found that trichome prevented the Empoasca fabae from laying eggs and feeding on the leaves, suggesting that soybean varieties with high trichome density are more resistant to the potato leafhopper, while those with low trichome density are less resistant [41]. Lahiri et al. (2020) found there was no correlation between trichome density and Megacopta cribraria infestation [42]. Zhang et al. (2022) found that the alfalfa variety (Caoyuan No. 4) showed high thrips resistance (more non-glandular trichomes in both leaves and stems) but fewer glandular trichomes and more wax in stems [43]. According to Yadav et al. (2020), Bemisia tabaci density exhibited a positive correlation with the density of leaf non-glandular trichomes [44].

There are fewer reports related to the resistance mechanisms of leaf shape and size traits. In this study, we found that, the longer the length of soybean leaves, the larger the aspect ratio; the narrower the leaves, the lower the number of thrips and the stronger the insect resistance. Megharaj et al. (2016) reported that leaf area had a negative association with thrip incidence, and trichomes were found to deter thrips from moderately resistant chili genotypes [45]. In contrast to our findings, their results showed that leaf area was non-significantly but positively correlated to the damage index. Leaf length, area, and thickness were found not to play any role in conferring resistance against Thrips tabaci on 12 cotton (Gossypium hirsutum) genotypes/varieties [46]. However, in contrast to these findings, our results indicated that leaf area was non-significantly but positively correlated with the damage index. Additionally, leaf length, area, and thickness were not significant factors in conferring resistance against Scirtothrips dorsalis and Polyphagotarsonemus latus [34]. Biochemical substances play an important role in the resistance of insect-resistant varieties to thrips. For instance, when Megalurothrips usitatus feeds, it strongly induces the expression of genes related to luteolin synthesis in the resistant variety, which indicates that biochemical substances likely confer greater resistance to M. usitatus [47]. Additionally, the correlations between the morphological structures and genotypes of insect-resistant and susceptible soybean varieties warrant further exploration.

The values of soil–plant analytical development (SPAD) can be used to express the chlorophyll content in soybean leaves [48]. The consistent and highly significant positive correlation between the chlorophyll content and SPAD values of soybean leaves further confirmed the feasibility of expressing chlorophyll content in terms of SPAD values [49,50]. Pests tend to feed on yellow-green plants, i.e., the lower the SPAD content of the leaves and the lighter the color of the leaves, the less resistant they are to insects [51]. In this study, we also found that the population size of thrips in 41 soybean varieties was significantly negatively correlated with the leaf SPAD values; i.e., the higher the leaf SPAD value, the lower the number of thrips.

According to the thrips amount ratio method, the 41 soybean varieties were broadly categorized into six classes: high resistance (HR), medium resistance (MR), low resistance (LR), low susceptibility (LS), medium susceptibility (MS), and high susceptibility (HS). This method allows for the evaluation of the relative insect resistance of the soybean varieties without specifying the resistance and susceptibility classes of the varieties. Although the method is relatively time consuming and laborious for identification, the results are stable, both in years of large thrips occurrences and in years of relatively light occurrences. Although there are some differences in the comparative results of these evaluation methods, the overall trend tends to be the same. The thrips amount ratio method takes more factors into consideration and can evaluate soybean resistance to thrips in a more comprehensive way, and its evaluation results are more reliable.

In this study, the population dynamics of T. flavus and the morphological structure of 41 spring soybean varieties were investigated for only one year, and a preliminary study of the resistance of the 41 spring soybean varieties to T. flavus was conducted. More investigations and analyses of thrips population dynamics and spring soybean varieties are needed to find stable resistance traits for production applications; further research is required on the physical resistance mechanisms to thrips in soybeans, as well as chemical resistance mechanisms, such as the determination of nutrients, including proteins, amino acids, and soluble sugars in soybean leaves.

Thysanoptera, an important group of phytophagous insects, are known to feed on a variety of crops, including legumes [52,53]. Considering the continued expansion of soybean acreage in China, pressure from harmful thrips is likely to increase. During the soybean flowering stage, thrips can cause some damage, such as anther rupture, which can lead to pollen loss [4]. This is a common symptom of reduced flower pollination and ultimately lower yields. This analysis aimed to identify key morphological and physiological traits that contribute to the resistance characterization of the 41 spring soybean varieties, providing insights into their agronomic performance and potential resistance mechanisms. Although knowledge of thrips resistance in soybean varieties is limited, conducting research in this area is critical to developing innovative pest management strategies. In the future, we will also consider referring to other reports to present a more accurate method of insect resistance identification [14].

5. Conclusions

This study evaluated the resistance of 41 soybean varieties to T. flavus and identified key morphological traits—such as compound leaf shape, leaf size, surface humidity, trichome density, color, and leaf nitrogen content—that are closely associated with insect resistance. The population dynamics of T. flavus during the blooming period can serve as a reliable indicator for assessing resistance levels. Based on our findings, “no pod trichome” was classified as a thrips-resistant variety, while “Ji Nong 29” was identified as a thrips-sensitive variety, providing useful reference materials for soybean breeding programs in Northeast China. These findings highlight the morphological basis of resistance to T. flavus, although the underlying physiological and molecular mechanisms have yet to be elucidated. Future work should focus on identifying these mechanisms and studying plant–insect interactions in field conditions.