Journal of Applied Botany and Food Quality 94, 1 - 6 (2021), DOI:10.5073/JABFQ.2021.094.001 Justus Liebig University, Institute of Botany, Giessen, Germany Common garden versus common practice – Phenotypic changes in Trifolium pratense L. in response to repeated mowing Thomas Gross*#, Christina Magdalena Müller#, Annette Becker, Birgit Gemeinholzer, Volker Wissemann (Submitted: June 15, 2020; Accepted: November 30, 2020) Summary Plants react to various biotic and abiotic factors by adjusting their growth pattern. This process, called phenotypic plasticity, can also be observed in the agronomical important fodder plant Trifolium pratense L. (red clover) which is grown worldwide for its high pro- tein content and soil improving ability. After cutting it changes its growth pattern due to phenotypic plasticity and initiates a second growth phase. Here, we analyze the regrowth dynamics of red clover in response to grazing or mowing by recording the plant’s architecture, leaf morphology, and growth performance in different cutting experiments. As previous studies were limited in the number of individuals and carried out via common garden experiments, we analyzed the regrowth reaction of T. pratense under standard agricul- tural field conditions. T. pratense forms smaller and rounder leaflets after repeated cutting, compensated by the production of more bio- mass. Nitrogen contents were subjected to seasonal changes, rather than by changes through cutting. As middle to late cut plants had higher regrowth capacities and regrew more biomass than early cut plants, an optimal time point for biomass harvest can be suggested to maximize the total yield of red clover biomass. Keywords: red clover, mowing, phenotypic plasticity, leaf round- ness, biomass production Introduction Plants are exposed to various biotic and abiotic factors, which they cannot evade due to their sessile nature. Therefore, they adapt their growth behaviour to local environmental constraints leading to phe- notypic variation owing to their open blueprint (Domagalska and leyser, 2011; Teichmann and muhr, 2015; gasTal and lemaire, 2015; Prins and Verkaar, 1992). Detailed phenotypic plasticity responses due to environmental perturbations were observed in many different plant species like Trifolium repens L., Populus trichocar- pa Hook., and Pisum sativum L. (sTafsTrom, 1995; goulas et al., 2002; ryle et al., 1985; Wu and sTeTTler, 1998). A major factor inducing phenotypic plasticity is the loss of biomass through her- bivory or cutting. The Fabaceae T. repens reacts to cutting by pro- ducing more but smaller leaves compared to uncut plants (goulas et al., 2002). As T. repens is a common fodder plant this effect might be of agronomical importance, as repeated cutting could reduce the amount of harvested biomass (gasTal and lemaire, 2015; Prins and Verkaar, 1992; Briske and richarDs, 1995). A close relative to T. repens, Trifolium pratense L. (red clover), is used in many parts of the world as high-quality fodder plant due to its naturally high protein content, deep root systems, and high yield which give T. pratense preference over T. repens. It is grown in monocultures or in a grass mixture where it increases the nitrogen content of the surrounding plants (eriksen et al., 2012; fernanDez and WaremBourg, 1987; Beecher et al., 2015; DeWhursT, 2013; kleen et al., 2011). In addition, red clover produces high amounts of the enzyme polyphenol peroxidase which increases both nitro- gen uptake and polyunsaturated fatty acids by livestock (lee et al., 2006; WinTers and minchin, 2005) and reviewed in Van ransT et al. (2011). Furthermore, digestibility analyses have shown the ad- vantages of re-grown red clover biomass in terms of protein content and digestible fiber content. In addition to varying growing condi- tions throughout the season, the cutting regime of red clover has a great influence on the harvested biomass, with respect to number of cutting events and their time points (Wiersma et al., 1998). However, both factors are also dependent on the respective red clover culti- var and their perseverance (marshall et al., 2017; Wiersma et al., 1998). In a previous study, phenotypic plasticity of red clover during re- growth after cutting/mowing was analysed in a common garden ex- periment (herBerT et al., 2018). After cutting, red clover produces, similar to T. repens, more leaves than uncut plants but they are of the same size than before cutting. Only the length of newly formed petioles is reduced when compared to uncut plants. Taken together, cutting triggers a second growth phase in red clover and could there- by contribute to yield increase. However, this analysis was performed under artificial conditions as a common garden experiment in a greenhouse using only a limited number of plants and these were cut only once. We were interested in the growth changes of T. pratense grown under field conditions with multiple cutting events and how they differ from the changes observed in the common garden. In par- ticular, it is interesting to see whether the second growth phase after the cut also contributes to increased yields in field trials. Here, we investigated the native response of locally adapted wild red clover plants as used in herBerT et al. (2018) to cutting and how different mowing regimes influence leaf morphology. We hypothesized that field grown red clover plants will only par- tially reflect the reaction shown by the common garden greenhouse experiment from herBerT et al. (2018) as they are more prone to abiotic and biotic stresses. Thus, the questions arise (i) whether and how plants compensate cutting losses during regrowth and by this if the experiment of herBerT et al. (2018) can be validated under common practice conditions in a field experiment, (ii) if different mowing regimes influence the amount of harvested biomass of the natural population red clover, and (iii) whether nitrogen content, as a marker for protein content and fodder quality, differs in the leaflets after mowing. Material and methods The experimental fields were located at the Agricultural Research and Developmental Farm of the Justus-Liebig-University Giessen in Rauischholzhausen/Germany (50°45’55.9008’’N, 8°52’1.2648’’E). The red clover seeds are derived from the same wild populations in Saxony, Saxony-Anhalt, Thuringia, Thuringian-Forest, and Ucker- * Corresponding author # These authors contributed equally to this work 2 T. Gross, C.M. Müller, A. Becker, B. Gemeinholzer, V. Wissemann marck as described in herBerT et al. (2018), (Rieger Hofmann seed company; Blaufelden, Germany). We sowed the red clover in au- tumn 2016, the mowing started in 2017. The plots were 2 m wide and 3 m long. Twelve different mowing regimes (Tab. 1), with 3 plots per regime, were applied. For further analysis plots 1, 3, 4, 6, 7, 9, 10, and 12 were considered, especially plots 4, 7, and 10 as those were mown three times. Each mowing regime was carried out three times on randomly distributed plots (Fig. 1, Tab. 1). The measurements were taken 18 - 20 days after mowing to evade the first stress responses. For each treatment, plant height, number of inflorescences, length and width of five leaflets per plant of ten randomly chosen plants were measured from three plots, totalling 30 plants per treatment and 150 leaflets. To avoid boundary effects in morphological measure- ments only plants internal of a 50 cm strip around the edges of each tested plot were analysed. The following characters were recorded: weight of the Fresh Material (FM), height of the plants, width and length of one leaflet, from the latter two characters we determined the roundness and the leaflet area. To calculate the roundness we used leaflet length / leaflet width, the leaflet area was determined by (leaf length / 2) * (leaf width / 2) * π. For nitrogen content measurements, the harvested T. pratense material was dried after mowing at 90 °C. The nitrogen content was measured with CN analyser Vario MAX (Elementar Analysensysteme GmbH, Hanau, Germany) (data was uploaded to the BExIS database and is accessible under https://doi. org/10.25829/bexis.26586-5, https://doi.org/10.25829/bexis.26606-1, and https://doi.org/10.25829/bexis.26666-1). For the questioning and the statistical calculation, we grouped the different plots in classes together after the time of mowing (Tab. 1): early means the time be- tween 12.6 - 12.09, middle: 03.07 - 02.10, late 14.07 - 13.10 and later 31.07 - 17.10, regardless of time: all plot s̓ which are mown three times (4, 7, and 10). Plot 1 is the unmown control field. The statistical tests and graphics were conducted in R (r core Team, 2013) with the fol- lowing packages PMCMR (PohlerT, 2014) and ggplot2 (Wickham, 2009). We tested for normality with the Shapiro-Wilk test and vari- ance homogeneity was tested with Bartlett test, if both tests were true, we applied a one-way analysis of variance (ANOVA). The Tukey HSD post hoc test was used to identify significant ANOVA results. If the normality and/or the variance homogeneity was not true, the Kruskal-Wallis test was applied and the Posthoc-Kruskal-Dunn test to further analyse the significant results of the Kruskal-Wallis test. Additionally we tested 16 different generalized linear mixed mod- els (GLMMs) using Template Model Builder (TMB) (Brooks et al., 2017), for the impact of Classes (early, middle, late, and later) the Cut systems (1/2/3/no) and the Time in the year (Supplement Tab. 1). The final model was chosen by the Akaike information criterion (AIC; Supplement Tab. 1). After first model validation, we transformed the data for leaflet roundness (standardized) and leaflet area (log) (Supplement Fig. 1, 2 and 3). Tab. 1: Mowing regime and dates of measurement for each plot Plot classes mowing regime 1st mowing Date 2nd mowing Date 3rd mowing Date of measurement of measurement of measurement 1 (a-c) not mown every time every time every time 2 (a-c) early 1× mown middle of June 30.06.2017 12.06.2017 3 (a-c) early 2× mown middle of June 30.06.2017 middle of August 29.08.2017 12.06.2017 14.08.2017 4 (a-c) early 3× mown middle of June 30.06.2017 middle of August 29.08.2017 middle of September 03.10.2017 12.06.2017 14.08.2017 12.09.2017 5 (a-c) middle 1× mown beginning of July 20.07.2017 03.07.2017 6 (a-c) middle 2× mown beginning of July 20.07.2017 end of August 19.09.2017 03.07.2017 31.08.2017 7 (a-c) middle 3× mown beginning of July 20.07.2017 end of August 19.09.2017 beginning of October 17.10.2017 03.07.2017 31.08.2017 02.10.2017 8 (a-c) late 1× mown middle of July 31.07.2017 14.07.2017 9 (a-c) late 2× mown middle of July 31.07.2017 middle of September 03.10.2017 14.07.2017 12.09.2017 10 (a-c) late 3× mown middle of July 31.07.2017 middle of September 03.10.2017 middle of October 31.10.2017 14.07.2017 12.09.2017 13.10.2017 11 (a-c) later 1× mown end of July 18.08.2017 31.07.2017 12 (a-c) later 2× mown end of July 18.08.2017 beginning of October 17.10.2017 31.07.2017 02.10.2017 Fig. 1: Structure of the test field of Trifolium pratense. 1-12 the different mowing regimes, a-c the randomly distributed plots. Reaction of T. pratense to repeated mowing 3 Results and discussion For plant height and the number of inflorescences, we could show that with increasing mowing frequencies the plant height and the number of inflorescences decreased compared to the not mown con- trol group (Suppl. Fig. 4 A and B). While these results show nicely the cutting effect on the regrowth ability of the red clover plants, our investigation concentrated on leaflet roundness and the potential increase in leaflet area associated with mass build-up in response to repeated cutting. For this, we identified the effect of different mow- ing frequencies (no cut, 1. cut, 2. cut, 3. cut) and time of mowing in the year (classes: early, middle, late, later) on leaflet roundness, leaflet area and the amount of harvested biomass. For all plots, which were mown at least two times (3, 4, 6, 7, 9, 10, and 12), the roundness of leaflets increased after the 2nd cut P < 0.001 (no normality distribution and no variance homogeneity P < 0.001), compared to the unmown red clover plants (control plants, plot 1) only the 2nd cut had significantly increased leaflet roundness. The 1st cut and the control plants showed no changes in leaflet roundness (P = 0.82; Suppl. Fig. 5 A). For the roundness of leaflets after three cuts regardless of the classes in the plots 4, 7, and 10 the data are not normally distributed (P = 0.003) and there is no variance homogeneity (P = 0.018; Fig. 2 A), the plots within the different cuts were compared. The Kruskal-Wallis test detected significance in the roundness of leaflets (P < 0.001) after cutting and the Posthoc-Kruskal-Dunn test identified that there are significant differences between all three cuts with P < 0.001. In the three plots 4, 7 and 10, differences were detected between plot 4 and the two other plots (7 and 10) in leaflet roundness (P < 0.001; Fig. 2 A). Compared to the control group (plot 1) there are no significant differences between the plots in the 1st cut and the control group (P = 0.129) in leaflet roundness. After the 1st cut, the roundness of the leaf- lets varied between 0.4 - 0.6, 2th cut 0.57 - 0.77 and 3rd cut 0.57 - 0.87 (with 1 being round; Fig. 2 A). The leaflet roundness increased with each cutting. Especially the plots 7 and 10 showed a great increase after the 2nd and 3rd cut (P = 0.01, Fig. 2 A and 3 B, C). Considering the four mowing classes (early, middle, late, and later), we observed a significant (P < 0.001) increase in leaflet roundness in all of our classes, regardless of the time period of mowing (Fig. 3 A-D). For plot 4 the roundness of leaflets after mowing differed sig- nificantly (P < 0.001) from the plots 7 and 10 (Fig. 2 A). We detected a significant difference in the roundness of leaflets in plot 4 between the 1st / 2nd (P = 0.002) and 1st / 3rd (P < 0.001) mowing (Fig. 3 A). Between the 2nd and 3rd mowing there were no differences detectable in Plot 4 (P = 0.49; Fig. 3 A). Our model (LM5) and the data show that the production of rounder leaflets after mowing is enhanced through multiple mowing events (Cut and Date; chi < 0.001, P < 0.001) and independent of the season. As herBerT et al. (2018) mowed red clover only once, this increase in roundness was not detectable. Leaf shape is often subjected to phenotypic plasticity depending on nutrient or water supply, and on temperature (li et al., 2019; Dorken and BarreTT, 2004; royer et al., 2009). However, leaflet roundness increases regardless of the classes, indicating a low or no contribution of these abiotic factors to the change in roundness. Interestingly, leaf shape is also dependent on the age of the plant, and in Glycine max (L.) Merr. older plants form less round leaflets (yoshikaWa et al., 2013). Hence, we can pro- pose that cutting rejuvenates T. pratense and early regrowth produces juvenile, rounder, leaflets. The leaflet area did not change for plots mown twice (3, 4, 6, 7, 9, 10, and 12) P = 0.160 (no normality distribution and no variance homogeneity P < 0.001) and only compared to the control plants the 1st and 2nd cut showed a significant decrease in the leaflet area (P < 0.001; Suppl. Fig. 1 B). After three cuts, regardless of the classes in the plots 4, 7, and 10, the leaflet area data is not normally distributed (P < 0.001) and there is no variance homogeneity (P = 0.026; Fig. 2 B). Significance was detected by the Kruskal-Wallis test in the leaf- Fig. 2: Development of (A) roundness of leaflets after mowing three times (up to 1 for round), (B) leaflet area [cm²] after mowing three times, (C) weight [kg_FM/m2] of the harvested plots after mowing three times in the year regardless of the season and (D) the nitrogen content [%] in the leaves after mowing. Red dots = outlier; Plot: 1 = not mown, 4 = early, 7 = middle, 10 = late. 4 T. Gross, C.M. Müller, A. Becker, B. Gemeinholzer, V. Wissemann let area (P < 0.001) after cutting and the Posthoc-Kruskal-Dunn test identified significant differences between all three cuts in terms of leaflet area with P < 0.001. Differences in leaflet area were detected between plot 4 and the two other plots (7 and 10, P < 0.001; Fig. 2 B). Compared to the control group (plot 1) there are no significant dif- ferences between no cut and 2nd cut (P = 0.399). After first mowing, the leaflet area varied between 4 - 5 cm² and remained stable after the second mowing (Fig. 2 B). Only after the third cut, the leaflet area of T. pratense decreases (Fig. 2 B). The Model LM4 shows a significant influence of mowing frequencies (chi < 0.001, P < 0.001) and the time of mowing (chi < 0.001, P < 0.001) for leaf let area. Considering the four treatment classes (early, middle, late, and later), we observed an increase of leaflet area after the 2nd cut and a de- crease after the 3rd cut in which it returns to the size of the 1st cut (early P < 0.001) or below (middle, late, later P < 0.001; Fig. 2 B). For plot 4 the leaflet area after mowing differed significantly (P < 0.001) from the plots 7 and 10 between the second and third mow- ing. The leaflet area in plots 7 and 10 decreased slightly after the 3rd cut (Fig. 2B), the leaflet area in plot 4 is relatively stable during the three times of mowing. Thus, the leaflet area of T. pratense remains rather stable even after two mowing events. This is in contrast to the reaction described in herBerT et al. (2018) and in T. repens (goulas et al., 2002), in which the leaflet area decreased already after the first cut. Red clover shows greater regrowth capabilities under field condi- tions than in common garden greenhouse conditions. For all plots mown twice (3, 4, 6, 7, 9, 10, and 12), we observed a sig- nificant increase in weight with 46.02 % (P < 0.001) (Suppl. Fig. 5 C). The biomass weight over three cuts regardless of the season was normally distributed (P = 0.083) and variance homogeneity was de- tected (P = 0.206). An ANOVA showed highly significant weight differences between the biomasses of different mowing regimes and the Tukey HSD post hoc test showed highly significant weight differ- ences among / between all three cuttings (Fig. 2 C) but not among the three plots (4, 7 and 10; P = 0.997). After the first cut, 8.18 kg Fresh Material (FM)/m² were harvested in total (Fig. 2 C, Tab. 2). Through the second time of mowing 11.97 kg FM/m² were harvested (Fig. 2 C, Tab. 2), thus the plants produced 46.38 % more above ground bio- mass after the first cut, a result similar to observations by eriksen et al. (2012) in a ryegrass / clover mixture. eriksen et al. (2012) also showed that a third cut reduced yield / weight of red clover. These results described here from the field experiments are in concordance with the common garden greenhouse experiments of herBerT et al. (2018) after which red clover initiates a new growth phase after cutting to compensate the loss of biomass. Similarly, this was also observed for the relative T. repens, which compensates the leaf loss Tab. 2: Weight development after three times of mowing regardless of the classes (only aspect the three cuts) and dependent on the class (early, middle, late, and later). Weight measure in kg fresh material (FM) per m². mowing Plots weight difference % kg_FM/m² in weight kg_FM/m² regardless of time 1. Cut 4, 7, 10 8.18 2. Cut 4, 7, 10 11.97 3.79 46.38 3. Cut 4, 7, 10 1.81 -10.16 -84.87 early 1. Cut 4 3.05 2. Cut 4 3.14 0.09 3.01 3. Cut 4 1.01 -2.13 -67.90 middle 1. Cut 7 2.68 2. Cut 7 4.09 1.42 52.96 3. Cut 7 0.59 -3.50 -85.54 late 1. Cut 10 2.45 2. Cut 10 4.73 2.28 93.20 3. Cut 10 0.21 -4.52 -95.56 later 1. Cut 12 2.38 2. Cut 12 3.45 1.08 45.26 Fig. 3: Development of leaflet roundness (up to 1 for round) (A) after early, (B) middle, (C) late and (D) later mowing in the year. ** P < 0.01, *** P < 0.001 Reaction of T. pratense to repeated mowing 5 after defoliation in 5-9 days (ryle et al., 1985; Black et al., 2009). However, repeated cutting seems to decrease the regeneration abili- ty of red clover (eriksen et al., 2012). These results are based only on the “three times mowing per year” regime regardless of mow- ing date. However, analysing the dates of mowing more thoroughly (Tab. 2, Fig. 3), suggests that mowing twice in the middle and late vegetation period is more beneficial for the gain of biomass weight, similar results for different cutting regimes were shown by Wiersma et al. (1998). An increase of weight of 50% - 90% was observed after the second cut, but a reduction in biomass of -84.87% as compared to the first cutting after the third cutting. The later cutting regime (only two cuts) improved yield to 45.26% (Tab. 2). However, it should be noted that there were 60 days between the 1st and 2nd cut and only 30 days between the 2nd and 3rd cut. However, the daily growth rate between the 1st and 2nd cut is 0.02 kg FM/m²/day, 0.073 times higher (early, middle, late) than between the 2nd and 3rd cut, thus we can as- sume that even with 60 days between the 2nd and 3rd cut, the amount of harvested biomass would have been less than after the 1st cut. For the early mowing the harvested biomass weight remained the same between the 1st and 2nd cut (P = 0.921) and decreased after the 3rd cut, in the other three classes the weight increased after the 1st cut and decreased after the 2nd cut (Fig. 3 A - D). Temporal analysis of plot 4 revealed that an early mowing is less suitable for increasing growth (Fig. 2 C, 3 A), similar results for an early mowing in the season were shown by Wiersma et al. (1998) which resulted in an reduced yield. We detected a change in nitrogen content over the year in the differ- ent mowing regimes 2nd cut – 3rd cut (P < 0.001). In all three plots, the nitrogen content increased during the season. Plot 4 had the biggest increase of the nitrogen content from 3.76% to 5.39% (Fig. 3 D), while the increases in plots 7 and 10 were lower, 3.46% to 4.18% and 3.77% to 4.46%, respectively. Based on the differences between plot 4 to 7 (P = 0.03; Fig. 3 D), mowing neither increased nor decreased the nitrogen content in our red clover plants. Dahlin and mårTensson (2008) could show a similar effect for T. repens after mowing. However, they used potted plants and mowed only once, thus the seasonal effect is missing. Nevertheless, even repeated cutt- ing did not reduce the nitrogen content, such that high protein content and quality of the harvested material remains stable. The perseverance of red clover was not a primary target of analy- sis in this study, but we found the absence of cutting negatively af- fecting the survival of red clover plants. After the eighth month, the red clover had died on the unmown control plots while mown red clover plots were healthy (Suppl. Fig. 6 and https://doi.org/10.25829/ bexis.26586-5). Conclusion In summary, we could validate the findings of the common garden greenhouse experiment from herBerT et al. (2018) into the field. T. pratense compensates the losses after cutting through starting a new growth phase, especially when the cutting was in the middle of June leading to fresh weight gain. Different mowing regimes influ- ence the harvested biomass: mowing at a later time in the year and after two cuts plants produce the highest biomass in the field. How- ever, the timing of the mowing had no influence on the N-content. Mowing early in the season does not significantly increase the bio- mass but mowing in the middle or later in the vegetation period (beginning of June) increases the biomass of the harvested red clo- ver and by this an increased total nitrogen content was achieved. However, even very late mowing (end of July) is more profitable than mowing three times early in the season (Tab. 2). Acknowledgements We like to thank Prof. Rod Snowdon and Lothar Behle-Schalk from the field station Rauischholzhausen and their team for technical sup- port with the field work. The help of Prof. Hans-Werner Koyro and Nicol Strasilla with the determination of the N-content is gratefully acknowledged. Also, we like to thank the students Dennis Oderwald, Larissa Witzel, and Julian Garrecht for assisting with the field measurements. This project was funded by the German Research Foundation (DFG Wi2028/8-2, BE2547/12-2, and GE1242/14-2). Conflict of interest No potential conflict of interest was reported by the authors. References Beecher, m., hennessy, D., BolanD, T.m., mceVoy, m., o’DonoVan, m., leWis, e., 2015: The variation in morphology of perennial ryegrass cultivars throughout the grazing season and effects on organic matter digestibility. Grass Forage Sci. 70, 19-29. DOI: 10.1111/gfs.12081 Black, a., laiDlaW, a., mooT, D., o’kiely, P., 2009: Comparative growth and management of white and red clovers. Ir. J. Agric. Food Res. 48, 149-166. Briske, D., richarDs, J., 1995: Plant responses to defoliation: a physiolo- gical, morphological and demographic evaluation. In: Bedunah, D.J., Sosebee, R.E. (eds.), Physiological geology and developmental morpho- logy, 635-710. Denver, Colorado: Society for Range Management. Brooks, m., krisTensen, k., BenThem, k., magnusson, a., Berg, c., nielsen, a., skaug, h., mächler, m., Bolker, B., 2017: glmmTMB Balances Speed and Flexibility Among Packages for Zero-inflated Generalized Linear Mixed Modeling. The R Journal 9, 378. DOI: 10.32614/RJ-2017-066 Dahlin, a.s., mårTensson, a.m., 2008: Cutting regime determines allo- cation of fixed nitrogen in white clover. Biol. Fertil. Soils 45, 199-204. DOI: 10.1007/s00374-008-0328-9 DeWhursT, r., 2013: Milk production from silage. Comparison of grass, legume and maize silages and their mixtures. Agric. Food Sci. 22, 57-69. DOI: 10.23986/afsci.6673 Domagalska, m.a., leyser, o., 2011: Signal integration in the control of shoot branching. Nat. Rev. Mol. Cell Biol. 12, 211-221. DOI: 10.1038/nrm3088 Dorken, m.e., BarreTT, s.c.h., 2004: Phenotypic plasticity of vegeta- tive and reproductive traits in monoecious and dioecious populations of Sagittaria latifolia (Alismataceae): a clonal aquatic plant. J. Ecol. 92, 32-44. DOI: 10.1111/j.1365-2745.2004.00857.x eriksen, J., askegaarD, m., søegaarD, k., 2012: Complementary effects of red clover inclusion in ryegrass-white clover swards for grazing and cutting. Grass Forage Sci. 69, 241-250. DOI: 10.1111/gfs.12025 fernanDez, m.P., WaremBourg, f.r., 1987: Distribution and utilization of assimilated carbon in red clover during the first year of vegetation. Plant Soil 97, 131-143. DOI: 10.1007/BF02149831 gasTal, f., lemaire, g., 2015: Defoliation, shoot plasticity, sward structure and herbage utilization in pasture. Review of the Underlying Ecophysio- logical Processes. Agriculture 5, 1146-1171. DOI: 10.3390/agriculture5041146 goulas, e., le Dily, f., simon, J.-c., ourry, a., 2002: Morphological pattern of development affects the contribution of nitrogen reserves to regrowth of defoliated white clover (Trifolium repens L.). J. Exp. Bot. 53, 1941-1948. DOI: 10.1093/jxb/erf045 herBerT, D.B., ekschmiTT, k., Wissemann, V., Becker, a., 2018: Cutting reduces variation in biomass production of forage crops and allows low- performers to catch up. A case study of Trifolium pratense L. (red clo- ver). Plant Biol. 20, 465-473. Stuttgart, Germany. DOI: 10.1111/plb.12695 kleen, J., TauBe, f., gierus, m., 2011: Agronomic performance and nutri- tive value of forage legumes in binary mixtures with perennial ryegrass under different defoliation systems. J. Agric. Sci. 149, 73-84. DOI: 10.1017/S0021859610000456 http://dx.doi.org/10.1111/gfs.12081 http://dx.doi.org/10.32614/RJ-2017-066 http://dx.doi.org/10.1007/s00374-008-0328-9 http://dx.doi.org/10.23986/afsci.6673 http://dx.doi.org/10.1038/nrm3088 http://dx.doi.org/10.1111/j.1365-2745.2004.00857.x http://dx.doi.org/10.1111/gfs.12025 http://dx.doi.org/10.1007/BF02149831 http://dx.doi.org/10.3390/agriculture5041146 http://dx.doi.org/10.1093/jxb/erf045 http://dx.doi.org/10.1111/plb.12695 http://dx.doi.org/10.1017/S0021859610000456 6 T. Gross, C.M. Müller, A. Becker, B. Gemeinholzer, V. Wissemann lee, m.r.f., colmenero, J.D.J.o., WinTers, a.l., scollan, n.D., minchin, f.r., 2006: Polyphenol oxidase activity in grass and its effect on plant- mediated lipolysis and proteolysis of Dactylis glomerata (cocksfoot) in a simulated rumen environment. J. Sci. Food Agric. 86, 1503-1511. li, g., hu, s., hou, h., kimura, s., 2019: Heterophylly: Phenotypic plas- ticity of leaf shape in aquatic and amphibious plants. Plants 8. DOI: 10.3390/plants8100420 marshall, a.h., collins, r.P., Vale, J., loWe, m., 2017: Improved per- sistence of red clover (Trifolium pratense L.) increases the protein sup- plied by red clover/grass swards grown over four harvest years. Eur. J. Agron. 89, 38-45. DOI: 10.1016/j.eja.2017.06.006 PohlerT, T., 2014: The Pairwise Multiple Comparison of Mean Ranks Package (PMCMR). R package. Prins, a., Verkaar, h., 1992: Defoliation: do physiological and morpho- logical responses lead to (over) compensation? In: Ayres, P.G. (ed.), Pests and Pathogens. Plant responses to foliar attack, 13-31. Oxford: Bios Scientific Publishers. r core Team, 2013: R: A language and environment for statistical comput- ing. R Foundation for Statistical Computing, Vienna, Austria. royer, D.l., meyerson, l.a., roBerTson, k.m., aDams, J.m., 2009: Phe- notypic plasticity of leaf shape along a temperature gradient in Acer ru- brum. PloS one 4, e7653. DOI: 10.1371/journal.pone.0007653 ryle, g.J.a., PoWell, c.e., gorDon, a.J., 1985: Defoliation in white clover: regrowth, photosynthesis and N2 fixation. Ann. Bot. 9-18. sTafsTrom, J., 1995: Influence of bud position and plant ontogeny on the morphology of branch shoots in pea (Pisum sativum L. cv. Alaska). Ann. Bot. 76, 343-348. DOI: 10.1006/anbo.1995.1106 Teichmann, T., muhr, m., 2015: Shaping plant architecture. Front. Plant Sci. 6, 233. DOI: 10.3389/fpls.2015.00233 Van ransT, g., lee, m.r.f., fieVez, V., 2011: Red clover polyphenol oxi- dase and lipid metabolism. Animal 5, 512-521. DOI: 10.1017/S1751731110002028 Wickham, h., 2009: ggplot2: elegant graphics for data analysis. New York, Springer-Verlag New York. Wiersma, D.W., smiTh, r.r., sharPee, D.k., mlynarek, m.J., ranD, r.e., unDersanDer, D.J., 1998: Harvest Management Effects on Red Clover Forage Yield, Quality, and Persistence. J. Prod. Agric. 11, 309-313. DOI: 10.2134/jpa1998.0309 WinTers, a.l., minchin, f.r., 2005: Modification of the Lowry assay to measure proteins and phenols in covalently bound complexes. Anal. Bio- chem. 346, 43-48. DOI: 10.1016/j.ab.2005.07.041 Wu, r., sTeTTler, r.f., 1998: Quantitative genetics of growth and develop- ment in Populus. III. Phenotypic plasticity of crown structure and func- tion. Heredity 81, 299-310. DOI: 10.1046/j.1365-2540.1998.00397.x yoshikaWa, T., ozaWa, s., senToku, n., iToh, J.-i., nagaTo, y., yokoi, s., 2013: Change of shoot architecture during juvenile-to-adult phase transi- tion in soybean. Planta 238, 229-237. DOI: 10.1007/s00425-013-1895-z ORCID Thomas Gross https://orcid.org/0000-0003-1194-6670 Christina Magdalena Müller https://orcid.org/0000-0003-4632-3166 Annette Becker https://orcid.org/0000-0002-3229-2162 Birgit Gemeinholzer https://orcid.org/0000-0002-9145-9284 Volker Wissemann https://orcid.org/0000-0003-0345-2603 Address of the corresponding author: Thomas Gross, Justus Liebig University, Institute of Botany, Heinrich-Buff- Ring 38, 35392 Giessen, Germany E-mail: Thomas.Gross@bot1.bio.uni-giessen.de © The Author(s) 2021. This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creative- commons.org/licenses/by/4.0/deed.en). http://dx.doi.org/10.3390/plants8100420 http://dx.doi.org/10.1016/j.eja.2017.06.006 http://dx.doi.org/10.1371/journal.pone.0007653 http://dx.doi.org/10.1006/anbo.1995.1106 http://dx.doi.org/10.3389/fpls.2015.00233 http://dx.doi.org/10.1017/S1751731110002028 http://dx.doi.org/10.2134/jpa1998.0309 http://dx.doi.org/10.1016/j.ab.2005.07.041 http://dx.doi.org/10.1046/j.1365-2540.1998.00397.x http://dx.doi.org/10.1007/s00425-013-1895-z I Supplementary material Common garden versus common practice – Phenotypic changes in Trifolium pratense L. in response to repeated mowing Short title: Reaction of T. pratense to repeated mowing Authors: Thomas Gross (1)*, Christina Magdalena Müller (1)*, Annette Becker (1), Birgit Gemeinholzer (1), Volker Wissemann (1) Supplement Table 1 Model selection after ACI. Roundness of leaflets~ (selected model orange), Leaf let Area ~ (selected model green), dF degrees of freedom, AIC Akaike information criterion, LogLik Log-Likelihood LEAF LET ROUNDNESS LEAF LET AREA MMOODDEELL Expression df AIC logLik df AIC logLik LLMM11 ~ Date 10 317.13519 -148.6 10 5045.236 -2512.6 LLMM22 ~ Classes 6 643.99998 -316 6 5642.2 -2512.6 LLMM33 ~ Classes + Cut 9 NA NA 9 5622.283 -2802.1 LLMM44 ~ Date + Cut 13 67.67025 -20.8 13 5008.606 -2491.3 LLMM55 ~ Date + Cut + (1| Classes) 14 65.11102 -18.6 14 5010.606 -2491.3 LLMM66 ~ Date + (1|Cut) 11 90.67765 -34.3 11 5023.724 -2500.9 LLMM77 ~ Date + (1|Cut) + (1| Classes) 12 80.65721 -28.3 12 5026.113 -2501.1 LLMM88 ~ Classes + (1|Cut) + (1|Date) 8 97.74024 -40.9 8 5054.068 -2519 LLMM99 ~ Date + (1|Patch) 11 102.66579 -40.3 11 5010.769 -2494.4 LLMM1100 ~ Classes + (1|Patch) 7 120.5131 -53.3 7 5066.344 -2526.2 LLMM1111 ~ Date + Classes + (1|Patch) 15 102.74644 -36.4 15 5010.808 -2490.4 LLMM1122 ~ Date + (1|Cut) + (1|Patch) 12 78.78658 -27.4 12 5011.891 -2493.9 LLMM1133 ~ Date + Classes + (1|Cut) + (1|Patch) 16 71.57755 -19.8 16 5012.808 -2490.4 LLMM1144 ~ Date + (1| Classes) + (1|Cut) +(1|Patch) 13 79.03038 -26.5 13 5013.891 -2493.9 LLMM1155 ~ Cut + (1|Date) +(1|Patch) 7 77.95899 -32 7 5032.99 -2509.5 LLMM1166 ~ Cut + Classes + (1|Date) +(1|Patch) 11 79.024 -28.5 11 5036.353 -2507.2 Supplement Figure 1: Plotted residual and fitted values from model LM5 leaflet roundness upper row and leaflet area lower row. A: without transformation of leaflet roundness; B: after transformation with standardize of leaflet roundness; C: without transformation of leaflet area. D: transformed log leaflet area. Supplementary material Supplement Tab. 1: Model selection after ACI. Roundness of leaflets~ (selected model orange), Leaflet Area ~ (selected model green), dF degrees of freedom, AIC Akaike information criterion, LogLik Log-Likelihood. Supplement Fig. 1: Plotted residual and fitted values from model LM5 leaflet roundness upper row and leaflet area lower row. A: without transformation of leaflet roundness; B: after transformation with standardize of leaflet roundness; C: without transformation of leaflet area; D: transformed log leaf let area. Supplementary material II Supplement Fig. 2: Model LM5 validation after transformation of leaflet roundness. A: Residuals vs. Cutting events; B: Residuals vs. experimental Day (Days of mowing events); C: Residuals vs. Classes (early, middle, late, later, and no~control); D: Roundness vs. Residuals in Cut events with smoothed re- gression line; E: Roundness vs. Residuals in Classes with smoothed regression line. Supplement Fig. 3: Model LM4 validation after transformation of leaflet Area. A: Residuals vs. Cutting events; B: Residuals vs. experimental Day (Days of mowing events); C: Residuals vs. Classes (early, middle, late, later and no~control); D: Roundness vs. Residuals in Cut events with smoothed regres- sion line; E: Roundness vs. Residuals in Classes with smoothed regression line. III Supplementary material Supplement Figure 3: Model LM4 validation after transformation of leaflet area. A: Residuals vs. Cutting events; B: Residuals vs. experimental day (days of mowing events); C: Residuals vs. Classes (early, middle, late, later and no~control); D: Roundness vs. Residuals in Cut events with smoothed regression line, E Roundness vs. Residuals in Classes with smoothed regression line. Supplement Figure 4: Trifolium pratense after the different mowing regimes. A: Height of plants; B: Number of inflorescences. Supplement Fig. 4: Trifolium pratense after the different mowing regimes. A: Height of plants; B: Number of inflorescences. Supplement Figure 5: Trifollium pratense after two times of mowing (regardless of the time) and the not mown control group. A: Roundness of leaflets in the different plots; B: Leaflet area in the different plots; C: Weight of the harvested biomass of the plots. Supplement Figure 6: Survival rate of the control red clover plants during the field experiment. Supplement Fig. 5: Trifolium pratense after two times of mowing (regardless of the time) and the not mown control group. A: Roundness of leaflets in the different plots; B: leaflet area in the different plots; C: weight of the harvested biomass of the plots. Supplement Figure 5: Trifollium pratense after two times of mowing (regardless of the time) and the not mown control group. A: Roundness of leaflets in the different plots; B: Leaflet area in the different plots; C: Weight of the harvested biomass of the plots. Supplement Figure 6: Survival rate of the control red clover plants during the field experiment. Supplement Fig. 6: Survival rate of the control red clover plants during the field experiment. Supplementary material IV