Based on a preliminary experiment, six varieties of herbaceous peony (Paeonia lactiflora Pall.) with strong, medium and weak waterlogging resistance and high ornamental value were selected: “Lihong”, “Dafugui”, “Hongxiuqiu”, “Qihualushuang”, “Taohuafeixue”, and “Yangfeichuyu”. Three-year-old ramets, healthy and disease-free, with strong growth were selected as experimental materials. After the vegetative growth of each variety was completed, field experiments were carried out (the cultivation substrate was garden soil).

The experiment was carried out from April 2019 to April 2020 at the Garden Center Laboratory of Horticultural Experiment Station of Shandong Agricultural University, the Gardening Center Laboratory and the Life Science Center, the Herbaceous Peony of the Horticulture Test Station and the Experimental Center of the Academy of Life Sciences. It is located between 116˚02'E to 117˚59'E and 35˚38'N to 36˚28'N, and is a temperate semi humid continental monsoon climate.

The experiment was divided into a waterlogging group (WL) and a control check group (CK), each treatment was repeated 3 times, 20 pots were used for sampling, and 5 pots were used for morphological observation. Waterlogging stress was removed when 60% (more than 60% is difficult to recover) of the leaves showed symptoms of injury. Waterlogging stress stage: It starts from the beginning of waterlogging stress treatment and ends when 60% of the leaves of each variety have symptoms of damage. Recovery stage: It begins when waterlogging stress is removed and ends when the leaves return to green (if it has not recovered in 18 days, it means that it cannot return to normal.). The WL used the “double-set basin” semi-flood method. Put the potted peony in a basin and keep the water level at 1/2 of the flowerpot to simulate waterlogging stress (flowerpot specification: 30 cm high, 33 cm upper diameter, and 27 cm lower diameter; basin specification: 41 cm long, 33 cm wide, and 20 cm high). Water the potted plants daily to keep the soil water supersaturated. The spacing of potted plants is 10cm. The reed curtains are covered on the surface of the pot to prevent the sun from increasing the water temperature in the pot. CK is managed according to conventional planting.

Photographs were taken every day to record growth status. Samples were taken every 3 days. Photosynthetic and chlorophyll fluorescence indices were measured every 6 days. Calculations of waterlogging tolerance and index screening formulas: The waterlogging stress phase is based on the stress time experienced by the earliest variety that was removed from the waterlogging stress, and the recovery phase is based on the time required for the earliest variety to be restored after the stress is removed. The index values of the tested varieties at the same time interval were selected to calculate and compare the waterlogging tolerance and recovery coefficients among varieties.

The chlorophyll content and relative conductivity of leaves were compared using the methods described by Zhao and Cang [9].

Remove the experimental materials from the main vein, weigh about 0.2 g of leaves into a 50 mL centrifuge tube, add 25 mL of 95% ethanol and seal it in sequence, extract at room temperature for 48 - 72 h under dark conditions, and subject the extract to UV754N ultraviolet-visible spectrophotometry Meter: Colorimetric comparison was performed at wavelengths of 665, 649, and 470 nm respectively. Calculate the concentration of chlorophyll a, b and carotenoid (mg/L) based on the OD value of each wavelength, and calculate the content of each pigment (expressed in mg/g fresh weight) based on the calculated concentration.

Ca = 13.95D₆₆₅ − 688D₆₄₉; Cb = 24.96D₆₄₉ − 732D₆₆₅; Cx = 1000D₄₇₀ − 205Ca − 114.8Cb/245.

The content of pigment in leaves (mg/g = pigment concentration (mg/L) x total volume of extract (mL) x dilution factor/sample mass (g).

Wash the fresh leaves with deionized water, punch with a hole punch with a diameter of 1 cm, take 0.1 g of the sample into a test tube, add 20 mL of deionized water, shake the shaker (230 r·min⁻¹), and determine the initial solution conductivity C₁, And then boil the water bath for 30 minutes, measure the conductivity C₂, and calculate REC = C₁/C₂ × 100% according to the formula.

The relative water content of the leaves was determined by the dry weighing method [10]. Cut the leaves, dry the surface, quickly weigh the fresh weight (FW), then immerse it in distilled water, place it in a dark place at 4˚C for 24 hours to absorb water, take it out, dry the surface water with absorbent paper, and weigh the saturated fresh weight (SFW), put them in an oven at 105˚C for 30 minutes, and dry them at 80˚C for 10 hours to constant weight. Weigh the dry weight (DW), and use the formula RWC = (FW-DW)/(SFW-DW) × 100% to calculate the relative moisture content.

Photosynthetic and chlorophyll fluorescence parameters were measured from 8:00 to 9:00 am on the sunny and windless days, using the 4^th to 5^th mature leaves from the parietal lobe. Photosynthetic gas exchange parameters were measured using a CIRAS-3 portable photosynthetic apparatus (PP Systems, USA). Chlorophyll fluorescence parameters were determined using a FMS-2 portable pulse-modulated fluorometer (Hansatech, UK).

Lutein-related indicators were quantified by HPLC. Conversion between the constituents of the lutein circulating component is expressed as (A + Z)/(A + Z + V) [11], where, A is the anther yellow matter content, Z is the zeaxanthin content, and V is the violaxanthin content.

Samples were taken from leaves at three stages: stress initiation, stress relief,and final recovery. The materials were fixed with FAA fixative solution and paraffin sections were made with reference to a previously described method [12]. Then, samples were observed under a Motic BA300 microscope, photographed with Motic Images Advanced 3.2 and Moticam 2006 to compare their structural changes, and microstructural parameters of leaves were measured using Image-Pro plus software. (1) Cell tense ratio (CTR, %) = the thickness of palisade organization/the thickness of leaf × 100; Spongy ratio (SR, %) = the thickness of sponge tissue/the thickness of leaf × 100; Vein protuberant degree (VPD, %) = the thickness of the vein/the thickness of leaf × 100

Data were processed and mapped using Microsoft Excel software, and data were tested for significance using DPS (Duncan’s new complex range method, P < 0.05). SPSS Statistics V 25.0 software was used for data principal component analysis, cluster analysis (membership function method) and stepwise regression analysis.

As shown in Table 1 and Figure 1(a), the stress tolerance time of different varieties varies from the beginning of waterlogging stress to the time when 60% of leaves turn yellow or red. “Dafugui” has the longest time (27 days), “Hongxiuqiu” and “Qihualushuang” have the shortest time (6 days). After the removal of stress, the recovery time of “Dafugui” was the longest (18 days) and “Taohuafeixue” was the shortest (6 days). When the stress was removed, the cell gap of

palisade tissue and sponge tissue of “Hongxiuqiu” increased significantly (Figure 1(a): Lms-H2, h2).

As shown in Figure 1(b) and Figure 1(c), compared with CK, the cell tense ratio (CTR) “Da fugui” increased the most (38.28%) when stress was removed, while the CTR of “Yangfeichuyu” and “Taohuafeixue” decreased the most (25.09% and 24.85%). During recovery, “Taohuafeixue” increased the most than CK (21.57%), while “Yangfeichuyu” decreased the most than CK (35.12%). The spongy ratio (SR) and CTR were negatively correlated. As shown in Figure 1(d), compared with CK, the ratio of “Taohuafeixue” palise tissue/sponge tissue (P/T) decreased the most (37.22%) and the ratio of “Da fugui” increased the most (64.31%) when the stress was removed. When recovering, the P/T of “Hongxiuqiu” increased the most (112.06%) and the P/T of “Yangfeichuyu” decreased the most (43.55%) compared with CK.

As shown in Figure 1(e), compared with CK, when the stress was removed, the leaf thickness of “Lihong” decreased the most, while that of “Hongxiuqiu” increased the most. At the time of recovery, “Yangfeichuyu” increased by 49.73% compared with CK. Vein protuberant degree (VPD) also changed significantly (Figure 1(f)). Compared with CK, the “great wealth” increased the most (20.48%) when the stress was removed, and the “red hydrangea” was slightly lower. During the recovery, “Lihong” and “Dafugui” increased the most (29.90 and 21.06%), while “Yangfeichuyu” decreased the most (30.47%).

According to Figure 2, the relative water content (RWC) of “Lihong” and “Yangfeichuyu” was significantly higher than that of CK when the stress was removed. When recovering, the RWC of “Dafugui”, “Taohuafeixue” and “Lihong” were significantly higher than those of CK. Other varieties are basically consistent with CK. The relative electrical conductivity (REC) represents the permeability of the cytoplasmic membrane. The larger the REC is, the more tissue exudate is, the more permeable the plasma membrane is, and the deeper the damage to the plant is [13]. Compared with CK, the REC of “Taohuafeixue” increased the most (27.04%) when the stress was removed, and the REC of “Yangfeichuyu” was slightly lower than CK. During recovery, the REC of “Taohuafeixue” increased significantly, 42.82% higher than CK.

Chlorophyll content is the direct manifestation of photosynthetic performance, nutritional status, and senescence in plant leaves. The external manifestation of waterlogging stress is the loss of green leaves, yellowing, and littering [14]. As shown in Figure 3, compared with CK, the Chla + b content in the leaves of

“Lihong” and “Dafugui” showed the greatest difference when the stress was removed, with a decrease of 46.03% and 46.24%, respectively. When recovering, “Lihong”, “Hongciuqiu”, “Qihualushuang”, “Taohuafeixue” all recovered, “Da fugui” was lower than CK, “Yangfeichuyu” was higher than CK. The main function of Chl b in plant leaves is to collect light energy, while the main function of Chl a is to convert the light energy collected by Chl b into chemical energy. The variation trends of Chl a and Chl b contents were consistent with that of Chl a + b. In the early stage of stress, Chl a/b was basically unchanged compared with CK. When the stress was removed, Chl a/b was lower than CK, and Chl a content decreased more.

As shown in Figure 4, following treatment, the photosynthetic rate (Pn), transpiration rate (E), and water utilization rate (WUE) of herbaceous peony varieties were significantly reduced by waterlogging stress. The Pn of “Lihong” was higher than CK in the 0 - 12 days of waterlogging stress, and increased 92.30% compared with CK on the 6^th day. When the stress was removed, “Lihong” and “Taohuafeixue” decreased the most than CK (94.05% and 94.66%). Compared with CK, intercellular CO₂ concentration (Ci) of the six varieties increased and stomatal

conductance (Gs) decreased when stress was removed.

The ratio of variable fluorescence (Fv) to maximum fluorescence (Fm) (Fv/Fm) represents the maximum photochemical efficiency of PSII and is a reliable indicator of the degree of photoinhibition [15]. When the stress was removed, the Fv/Fm of “Hongxiuqiu” and “Yangfeichuyu” decreased the least compared with CK, indicating a lower degree of photoinhibition. NPQ is the coefficient of heat dissipation. As shown in Figure 5(a), when the stress is removed, the values of “Dafugui”, “Taohuafeixue” and “Yangfeichuyu” increase the most than CK (205.66%, 170.27%, and 150.05%), which reflects its strong heat dissipation capacity and can better consume excessive energy through heat dissipation after stress.

qP represents photochemical quenching, which reflects the openness of the PSII reaction center. ETR is the apparent electron transfer rate, and the variation trend of ETR and qP in this experiment is basically the same. Compared with CK, the ETR of “Qihualushuang” increased the most (110.47%) and the ETR of “Dafugui” decreased the most (49.77%) when the stress was removed (Figure 5(b)). When recovering, the ETR of “Lihong” and “Dafugui” decreased, while that

of “Hongxiuqiu” increased, and other varieties basically recovered. When the stress was removed, the qP of “lihong” increased the most (43.28%), and that of “red hydrangea” decreased the most (32.46%). When recovering, the qP of “Lihong” and “Dafugui” decreased, and that of “Qihualushuang”, “Taohuafeixue”, and “Yangfeichuyu” increased.

Lutein comprises violaxanthin (V), anthe-raxanthin (A), and zeaxanthin (Z). In the xanthophyll cycle, the three components of lutein (A, Z, and V) are mutually transformed according to changes in light conditions, and both A and Z dissipate heat and rely on the xanthophyll cycle. The heat dissipation capacity is positively correlated with the contents of A and Z [16]. As shown in Figure 6, following waterlogging treatment, the (A + Z)/(A + Z + V) of each variety increased. Compared with CK, the changes in “Taohuafeixue”, “Yangfeichuyu”, and “Lihong” increased significantly by 140.86%, 211.16%, and 90.20%, respectively. It showed that waterlogging stress increased the conversion of V to A and Z, and that “Linhong”, “Taohuafeixue”, and “Yangfeichuyu” can effectively upregulated the lutein cycle to consume excess excitation energy, and thus better protect the light system from damage.

Stress resistance in plants is a complex quantitative trait that is affected by many factors. Different varieties have different mechanisms of waterlogging tolerance, and present different responses to specific indicators under adverse conditions [17] [18]. According to the methods and formulas described by Zhou et al. [17], the principal component analysis and membership function analysis were used to comprehensively analyze the tolerance and recovery degree of the tested varieties, and the indexes values of each variety at the 6^th day after waterlogging stress and the 6^th day after removal stress were selected to calculate and compare

the waterlogging resistance coefficient and recovery coefficient.

Table 2 shows the different variation ranges of the waterlogging resistance coefficient of each single index of the six varieties. Table 3 is the correlation coefficient matrix of each variety, which shows that there are large or small correlations among all physiological indexes. Through principal component analysis of the waterlogging resistance coefficient of 15 individual indexes, the cumulative contribution rate of the first four comprehensive indexes reached 95.94% (Table 4), and the contribution rate reflected the relative importance of each comprehensive index. According to the index coefficient of each comprehensive index (Table 4) and the waterlogging resistance coefficient of each single index (Table 2), the values of the four comprehensive indexes of each variety (Table 5) were calculated.

As shown in Table 5, the waterlogging tolerance D value of each variety was

ranked as follows: “Lihong” > “Yangfeichuyu” > “Taohuafeixue” > “Dafugui” > “Hongxiuqiu” > “Qihualushuang”. The D value for recovery degree of each variety was calculated by the same method, and the degree of recovery after the removal of stress was: “Qihualushuang” > “Lihong” > “Hongxiuqiu” > “Dafugui” > “Yangfeichuyu” > “Taohuafeixue”. After a short period of stress (6 days), symptoms of stress appeared; the D value of waterlogging tolerance was the minimum, the recovery process was the fastest and the strongest, indicating that the “Qihualushuang” was the most sensitive to waterlogging stress. The longest duration of stress (27 days) was experienced by “Dafugui”, with moderate waterlogging tolerance (D value) and slow recovery, indicating that “Dafugui” was less sensitive to waterlogging stress; the morphological changes in plants were not obvious in the early stage of stress, but they were physically damaged. The maximum D value of waterlogging resistance, and the rapid recovery of “Lihong” showed that higher water content in the early stage (about 12 days) was beneficial to its growth.