# Source paper archive - Source title: Drought Response and Recovery in Wild, Feral, and Domesticated Brassica rapa: Implications for Crop Climate Resilience - Author(s): Borsuk, A. M., Quartuccia, G. I., Talpins, C., Angert, G., Warner, R., Vega, L. R., Leahy, M., Mittal, R., Daybydeen, L. C., Grippo, I., Bichler, J., Moore, E. G., Salha, R., Ivanovski, K., Ellingboe, A., Mattalia, G., Mabry, M. E., Diazgranados, M., Franks, S. J., McAlvay, A. C. - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.06.764879v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.06.764879v1.full.pdf - PDF SHA-256: f5df3dc2afd50717f159bf960b847cfa63f337332d0fa8e094be6f9e850a8c1d - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) # 1 Drought Response and Recovery in Wild, Feral, and Domesticated Brassica rapa : 2 Implications for Crop Climate Resilience 3 Aleca M. Borsuk [1†], Gabriella I. Quartuccia [2†], Chase Talpins [1,3], Grace Angert [1,2], Rachel Warner [1,4], 4 Luis R. Vega [2], Meredith Leahy [2], Rhea Mittal [2], Leeann C. Daybydeen [1], Isabella Grippo [2], Julie 5 Bichler [1,5], Elisabeth Moore [1,5], Reem Salha [1,5], Kelly Ivanovski [1], Anna Ellingboe [1], Giulia Mattalia [6], 6 Makenzie Mabry [7], Mauricio Diazgranados [1], Steven J. Franks [2], Alex C. McAlvay [1 ] 7 1 New York Botanical Garden, Bronx, NY, USA, 2 Department of Biological Sciences, Fordham 8 University, Bronx, NY, USA, [3] Rye Country Day School, Rye, NY, USA, [4] Columbia College, 9 Columbia University, New York, NY, USA, [5] Barnard College of Columbia University, New York, 10 NY, USA, [6] Laboratori de Botànica–Unitat Associada CSIC, Facultat de Farmàcia i Ciències de 11 l’Alimentació–Institut de Recerca de la Biodiversitat (IRBio), Universitat de Barcelona, Barcelona, 12 Catalonia, Spain, [7] Florida Museum of Natural History, University of Florida, Gainesville, FL, USA 13 - These authors contributed equally to the work 14 **Corresponding Author** : Aleca Borsuk; aborsuk@nybg.org 15 **ORCIDs** : Aleca M. Borsuk: 0000-0002-1696-9647, Gabriella I. Quartuccia: 0000-0002-4377-4028, 16 Chase Talpins: 0009-0000-5679-8152, Grace Angert: NA, Rachel Warner: NA, Luis R. Vega: 0009 17 0006-3762-6754, Meredith Leahy: 0009-0007-4020-6416, Rhea Mittal: NA, Leeann C. Daybydeen: 18 0009-0001-6502-1668, Isabella Grippo: NA, Julie Bichler: 0009-0002-9102-3250, Elisabeth Moore: 19 0009-0007-9158-5225, Reem Salha: 0009-0008-5995-9482, Kelly Ivanovski: NA, Anna Ellingboe: 20 NA, Giulia Mattalia: 0000-0002-1947-7007, Makenzie Mabry: 0000-0002-6139-9559, Mauricio 21 Diazgranados: 0000-0003-0448-5706, Steven J. Franks: 0000-0001-9681-3038, Alex C. McAlvay: 22 0000-0001-7051-2018 23 **Short title** : Drought response in _Brassica rapa_ 24 **Keywords** : _Brassica rapa_, crop wild relatives, domestication, drought resilience, feral, 25 photosynthesis, taproot, water use efficiency 26 **Date of submission** : 11/6/2026 | **No. figures** : 4 | **Word count** : 4871 | **Supp. Figs** : 3, **Supp. Tables** : 6 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 28 **Highlight** 29 Drought-resilient traits and genotypes were identified in wild, feral, and domesticated accessions of 30 _Brassica rapa_, providing insight into evolution under domestication and feralization in the context of 31 climate change. 32 **Abstract** 33 Drought poses a serious threat to agricultural production and food security. Wild relatives of crops, 34 including pre-domesticated and feral varieties, contain genetic diversity that can serve as a breeding 35 resource for increasing desirable crop traits such as drought resilience, but little is known about the 36 relative ability of wild varieties to cope with climate change in most species. In this study, we 37 investigated the drought response of wild, feral, and domesticated varieties of _Brassica rapa_, an 38 important crop and model species. Twenty-five accessions were subjected to experimental drought 39 and recovery, and variation in traits such as biomass retention, photosynthetic rate, and water use 40 efficiency were quantified. We found several wild accessions of _B. rapa_ had high biomass retention 41 under drought and full recovery of maximum assimilation rate after rewatering, challenging the 42 assumption that feralization produces stress-resilient phenotypes and suggesting that populations of 43 wild crop progenitors are especially promising reservoirs of drought-response diversity. Our results 44 also showed taproot biomass was positively associated with water use efficiency. Our findings 45 provide candidate traits and genotypes for agricultural breeding aimed at increasing drought 46 resilience in domesticated _B. rapa_ cultivars and give insights into evolution under domestication and 47 feralization in the context of climate change. 48 **Introduction** 49 Drought is one of the most severe constraints on crop productivity worldwide, and its impacts are 50 becoming increasingly problematic due to climate change (Trenberth et al., 2014; Kogan et al., 2019; 51 Begna and Wakweya, 2025). Despite its importance to agriculture, drought resilience, i.e. a plant’s 52 ability to recover after experiencing drought stress, remains a challenging trait to understand and 53 breed for due to the fact that it involves multiple physiological, morphological, and developmental 54 processes (Khadka et al., 2020). These responses are context-dependent and often controlled by 55 multiple genes, making it difficult to select for drought-resilient genotypes using conventional 56 breeding approaches (Mir et al., 2012). [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 57 Crop wild relatives (CWRs) are potentially valuable resources for breeding crops with enhanced 58 drought resilience due to relatively high genetic diversity and traits that enable survival under harsh 59 environmental conditions (Zhang et al., 2017; Jovovic et al., 2020). CWRs offer reservoirs of genetic 60 diversity that can contribute to crop resilience under climate change (Dempewolf et al., 2014; 61 Brozynska et al., 2016). In recent years, interest has grown in leveraging these populations not only 62 as a source of traits but also through novel approaches such as _de novo_ domestication—where wild 63 organisms are domesticated from scratch (Fernie and Yan, 2019). Despite this potential, there is 64 limited information on how domesticated crops and their wild forms vary in terms of physiological 65 traits such as photosynthetic rate, water use efficiency, and stomatal conductance (Lei et al., 2024). 66 Feral crops, which are formerly domesticated populations of plants that have escaped cultivation and 67 form free-living populations, are also potentially underutilized pools of adaptive diversity for 68 breeding stress resilience, as these populations often persist in disturbed or marginal habitats, 69 suggesting they may possess valuable stress resilience traits (Ellstrand et al., 2010; Mabry et al., 70 2023). _De novo_ re-domestication, which uses feral populations as raw material to breed new crops, 71 has also been proposed as another way to leverage adaptive diversity (Pisias et al., 2022). Yet, 72 feralization is relatively understudied: while studies have demonstrated the presence of 73 “domestication syndromes,” which are morphological, physiological, and developmental traits that 74 commonly arise in groups of plants (and animals) under domestication (Meyer et al., 2012; Rendón 75 Anaya and Herrera-Estrella, 2018), analogous “feralization syndromes” are comparatively less well 76 characterized, especially outside of work on the primary model for feral plant biology, weedy rice 77 (Gressel, 2005; Mabry et al., 2021; Mabry et al., 2023). Pathways to feralization do not necessarily 78 follow the reverse of domestication, making it important to understand feral plant physiology and 79 genetics as unique phenomena (Niego and Benítez-Burraco, 2022). 80 The physiological mechanisms underlying drought responses remain poorly understood across 81 domestication status, i.e., from wild to domesticated to feral, in most crop systems (Milla et al., 82 2015). In particular, there is a lack of direct experimental studies that compare plant responses across 83 domestication status under controlled drought conditions, limiting our understanding of how 84 domestication and ferality shape stress adaptation and our ability to predict the presence of beneficial 85 traits in wild and feral crop relatives. 86 _Brassica rapa_ presents a powerful model system for understanding drought response from the 87 perspective of evolutionary ecology across domestication status. _B. rapa_ includes a range of crop [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 88 types selected for different target organs, including leafy vegetables (e.g., bok choy), root crops (e.g., 89 turnip), and oilseeds (e.g., yellow sarson). The genus _Brassica_ includes a number of crops of major 90 economic and nutritional importance, with many of them forming what is called the “Triangle of U” 91 with the diploid species _Brassica rapa_, _Brassica oleracea_, and _Brassica nigra_ on the vertices, and 92 allotetraploids of each combination of the three diploids forming the sides (Nagaharu and Nagaharu, 93 1935). However, _B. rapa_ in particular has emerged as a valuable model for multiple areas of plant 94 biology due to its phylogenetic proximity to the plant model _Arabidopsis thaliana_, genomic 95 resources, rapid cycling forms, and role as a genome donor to the well-studied allopolyploid relatives 96 _Brassica napus_ and _Brassica juncea_ (McAlvay et al., 2021). Drought poses a major constraint for 97 many _Brassica_ crops, with significant yield losses reported under water-limited conditions 98 (Chikkaputtaiah et al., 2017). Field mustard, _B. rapa_, has also emerged as an important model for 99 understanding rapid evolutionary responses to drought (Franks, 2011; Franks et al., 2016; Hamann et 100 al., 2018), including early flowering in wild populations as a mechanism for drought escape (Franks, 101 2011). 102 To date, comparative studies in _B. rapa_ across the entire wild-feral-domestication spectrum have 103 been hampered by a lack of clarity over which populations are feral vs. truly wild (McAlvay et al., 104 2017). In a recent advance, population genetic work has clarified the domestication status in 653 105 accessions of _B. rapa_, identifying wild lineages from the Caucasus, Siberia, Italy, and Algeria. The 106 study used genotyping-by-sequencing and habitat suitability mapping to understand genetic 107 relationships, diversity, admixture, and structure to distinguish their evolutionary status. This set of 108 resolved wild, feral, and domesticated populations provides an unusually clear system to compare 109 drought response across domestication status (McAlvay et al., 2021). 110 In this study, we took advantage of these newly resolved genetic relationships to understand the 111 impact of domestication status on _B. rapa_ response to drought. Specifically, we aimed to understand 112 if wild, domesticated, and feral _B. rapa_ differ in drought response, and if so, what physiological traits 113 support recovery from drought, i.e., resilience. To inform future crop breeding efforts, we also aimed 114 to understand if specific accessions of _B. rapa_ had elevated drought resilience. Finally, we sought to 115 understand if there was variation in drought response across crop types (e.g., leafy vs root vs oilseed), 116 including the impact of fleshy taproots on plant water use efficiency. To address these questions, we 117 carried out a greenhouse drought and recovery experiment comparing 25 accessions of _B. rapa_ across 118 domestication status and crop types, incorporating gas exchange, chlorophyll fluorescence, and [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 119 biomass data. The overall goal of this work was to contribute to our fundamental understanding of 120 adaptive drought response in wild, domesticated, and feral plants, and to inform agricultural breeding 121 efforts aimed at increasing drought resilience in _B. rapa_ crops. 122 **Materials and Methods** 123 **Plant material and growth conditions.** A set of 25 accessions of _B. rapa_ was selected to represent 124 a broad range of geographies, domestication statuses, and morphotypes. This included nine wild 125 accessions, eight feral accessions, and eight crop accessions (three turnips, two oilseeds, and three 126 ‘leafy’, i.e. bok choy/choy sum). All accessions were selected as a subset of total accessions 127 sequenced for a previous study (McAlvay et al., 2021) to identify their status as wild, feral, or the 128 specific crop. Fresh seeds were obtained from the Leibniz Institute of Plant Genetics and Crop Plant 129 Research, United States Department of Agriculture, and UK Vegetable Genebanks. Plants were 130 grown in 6-inch standard pots in a commercial high-porosity soil (PRO-MIX HP Mycorrhizae; 131 Premier Tech Horticulture, Québec, Canada) containing 65-75% sphagnum moss, perlite, limestone, 132 a wetting agent, a starter nutrient, and a mycorrhizal component ( _Rhizophagus irregularis_ ), amended 133 with compost and RootShield granular biofungicide ( _Trichoderma harzianum_ strain T-22) at the 134 Nolen Greenhouses at the New York Botanical Garden. 135 Potted plants were arranged in a randomized block design during the experimental phase (Fig. S1). 136 Each block consisted of one individual of each of the 25 accessions. A total of ten blocks (A through 137 J) were established, five of which were subjected to experimental drought and recovery (i.e., 138 ‘drought’ treatment) and five of which were consistently well-watered (i.e., ‘control’ treatment). 139 All plants were watered by hand to maintain soil hydration throughout germination and 140 establishment. Starting at the onset of the experimental phase, blocks A-E were designated as the 141 control treatment, and watered once per week by saturating the soil completely via timed (10-min) 142 drip irrigation. Blocks F-J were designated as the drought treatment and only watered in week 0 (start 143 of the experiment) and then again in week 3 (recovery watering), to full saturation of the soil via 144 timed (10-min) drip irrigation (Table S2). 145 **Gas exchange and chlorophyll fluorescence.** Leaf gas exchange was measured 1-3 days 146 immediately following plant watering (due to the number of replicates, gas exchange measurements 147 could not be completed on all plants in a single day). Rate of maximum CO 2 assimilation (A max ) was 148 measured using a LI-6800 portable gas exchange system (Licor Inc, Lincoln, NE, USA); leaves were [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 149 placed in the chamber during the hours of 9 am to 2 pm and allowed to acclimate to the following 150 conditions until they reached a steady state: PPFD = 800 µmol m [-2 ] s [-1], CO 2 = 400 ppm, RH = 37%, 151 VPD = 1.6 kPa, T leaf = 25 ˚C. The light intensity for all A max measurements was standardized at 800 152 PPFD, to reduce photoinhibition during drought and accommodate a variety of accessions adapted to 153 different light environments. Stomatal conductance to water (g sw ) was monitored simultaneously with 154 assimilation rate to calculate instantaneous water use efficiency (WUE = A max /g sw ). To estimate 155 photosynthetic efficiency as an indicator of stress, the quantum yield of photosystem II (YII) was 156 measured pre-dawn (between ~3 am to ~6 am) the day after watering on dark-adapted plants using a 157 MINI-PAM-II fluorometer (WALZ GmbH, Effeltrich, Germany). For all gas exchange and 158 chlorophyll fluorescence data, measurements were obtained from 1 leaf per individual and replicated 159 on 5 individuals per accession for each treatment (drought and control). Measurements were obtained 160 from leaves at the same developmental stage (leaf 4-5 in weeks 0-1, and leaf 5-6 in weeks 2-3). 161 **Soil moisture.** Soil moisture was tested synchronously with gas exchange measurements using a 162 Hydrosense II soil moisture sensor (Campbell Scientific, Logan, Utah) to ensure the drought 163 treatment was effective and comparable across domestication status and crop type (Fig. S2). The 164 moisture probe was inserted vertically at the midpoint between the stem and pot wall and held in 165 place until the measurement stabilized. For each sampling event, two measurements were taken at 166 opposite sides of the pot. The two volumetric water content readings were averaged to determine the 167 mean soil moisture. 168 **Biomass.** Plant biomass was collected for all individuals at the conclusion of the growth experiment 169 and partitioned into vegetative, hypocotyl (i.e., fleshy taproot), and root tissues. Roots were separated 170 from soil and other particulates by first submerging the root ball in water to remove bulk soil content, 171 then subsequent washing of the roots under running tap water. All tissue samples were dried in paper 172 bags at 60°C in a gravity oven (Fisherbrand™, Waltham, Massachusetts, USA). Dry weight of each 173 sample was weighed using a 0.001 g balance (Sartorius, Göttingen, Germany). 174 **Statistical Analysis.** 175 _Biomass retention._ Total plant biomass (the sum of vegetative, hypocotyl, and root tissue mass) was 176 calculated for each accession in drought vs. control treatments. Biomass retention was calculated for 177 each accession as the ratio of total plant biomass in the drought treatment to total plant biomass in the 178 control treatment (Biomass retention = Drought biomass / Control biomass). Values >1 indicate [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 179 relative biomass gain in the drought treatment compared to the control treatment, while values <1 180 indicate biomass loss in the drought treatment compared to the control treatment. 181 _Comparison of biomass retention group means._ Linear models were used to test the effect of 182 domestication status (wild, n =9; feral, n = 8; domesticated, n = 8) on biomass retention. Model 183 assumptions were checked with Shapiro-Wilk and Levene’s test. Estimated marginal means were 184 calculated via the _emmeans_ package in R (Lenth and Piaskowski, 2025) and pairwise comparisons 185 were run between categories using Šidák adjustment for multiple comparisons. 186 _Physiological response models._ Linear mixed effects models (LMEM) were run via the “lmer” 187 command from the _lme4_ package in R (Bates et al., 2015) to test how A max, YII, water use efficiency, 188 or stomatal conductance (physiological response; Tables S3-S5) varied with fixed and random 189 effects. LMEM included treatment, week, and domestication type as fixed effects with all two- and 190 three-way interactions between terms, as well as the random effects of block and plant ID nested 191 within accession. The _emmeans_ package was used to make pairwise comparisons between groups via 192 Tukey adjustment. LMEM were run to compare cultivar physiological response at the weekly level, 193 as well as across two weeks at a time to compare the rate of change in physiological response 194 between the two treatments. LMEM report Tukey-adjusted p-values. Linear models were additionally 195 run using “lm” regression from the _stats_ package in R (R Core Team, 2025) to investigate the 196 relationship between water use efficiency and mean taproot biomass. 197 198 **Results** 199 **Biomass Retention.** The drought treatment significantly reduced biomass overall, as expected, with 200 plants in the drought treatment having on average 78.6% of the biomass of plants in the watered 201 treatment (t (48) = -2.64, p = 0.011). However, there was substantial variation among varieties, with 202 some having greater biomass under drought than control conditions. Specifically, a biomass retention 203 index (drought biomass / control biomass) greater than 1, which indicates greater biomass in drought 204 than control conditions (overcompensation), occurred in three wild accessions native to the country 205 Georgia, a domesticated accession from Finland, and a feral accession from Guatemala (Fig. 1A; 206 Table S1). In contrast with these highly drought resilient varieties, some were very drought sensitive, 207 including two domesticated and one feral variety that had less than half of the biomass under drought 208 compared with control conditions (Fig. 1A). Comparing accessions by domestication status, biomass [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 209 retention was significantly higher in wild accessions (mean = 1.05 ± 0.089) than in domesticated 210 accessions (mean = 0.723 ± 0.089; t (22) = -2.60, p = 0.049), with feral plants (mean = 0.799 ± 0.101) 211 not differing from either wild (t (22) = -1.86, p = 0.212) or domesticated accessions (t (22) = -0.57, p = 212 0.924; Fig. 1B). When domesticated accessions were further compared by crop type (e.g., leafy vs 213 root vs oilseed), and after adjusting for multiple comparisons, no crop type differed significantly in 214 biomass retention (Fig. S3). 215 216 **Figure 1.** Biomass retention (drought biomass/control biomass) comparisons across 25 accessions of 217 domesticated, feral, and wild _Brassica rapa_ . (A) Ranking of accessions by biomass retention; dashed 218 line indicates a biomass retention value of one. (B) Comparison of biomass retention by 219 domestication status. Box upper and lower boundaries represent the 75 [th] and 25 [th] percentiles, 220 respectively, and lines across boxes represent the group median. Whiskers extend from the upper and 221 lower quartiles to the group maximum and minimum, respectively. Group means not sharing a letter 222 are significantly different by pairwise comparisons with Šidák adjustment (p < 0.05). Colors indicate 223 domestication status as shown in (A). 224 225 **Photosynthesis and Quantum Yield.** In week 0 (no difference between treatments; all plants were 226 well-watered), maximum photosynthetic rate, estimated as the steady-state rate of CO 2 assimilation at 227 800 PPFD (A max ), was significantly higher in wild accessions relative to domesticated accessions 228 (t (31.3) = -3.256, p = 0.007) according to Tukey-adjusted tests for pairwise comparisons, indicating 229 greater photosynthetic productivity in wild plants compared to domesticated plants under non 230 stressed conditions (Fig. 2A, Table S3). In weeks 1-2, when drought-treated plants did not receive 231 any water, A max was lower for plants in the drought treatment compared to the control treatment, and 232 there was no significant effect of domestication status (wild, feral, or domesticated) on the reduction [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 233 in A max due to drought (Fig. 2A; Table S3). After rewatering (week 3), wild plants in the drought 234 treatment had significantly greater recovery in A max compared to domesticated plants in the drought 235 treatment (t (60.5 ) = -3.13, p = 0.0076), indicating higher physiological resilience to multi-week 236 drought stress in wild accessions (Fig. 2A; Table S3). Further, in week 3, wild accessions that 237 underwent drought fully recovered maximum photosynthetic rates upon rewatering, with A max values 238 not significantly different from the same accessions in the control treatment (t (123) = 0.50, p = 0.616). 239 In comparison, A max in drought-stressed domesticated (t (117) = 2.22, p = 0.028) and feral (t (115) = 3.31, 240 p = 0.001) accessions remained depressed relative to the respective accessions in the control 241 treatment (Fig. 2A). 242 For plants in the drought treatment, the change in A max from week 0-1 (shown as the initial slope in 243 Fig. 2A) indicates the degree of sensitivity to drought, with a slope not significantly different from 244 the plants in the watered treatment indicating a lack of drought sensitivity. The change in A max from 245 week zero to one was significantly lower in the drought than control plants for wild (t (130) = 2.26, p = 246 0.025) and feral (t (123) = 2.25, p = 0.026) plants, but not for domesticated plants (t (132) = 1.65, p = 247 0.101), suggesting relatively lower sensitivity to drought in domesticated plants. However, it is 248 important to note that in week 1, A max was greater in wild and feral than in domesticated varieties, 249 with A max declining from week 0-1 in both watered and drought treatment domesticated plants (Fig. 250 2A). The change in A max from week one to two was significantly different between the watered and 251 drought treatment for wild plants (t (128) = 5.74, p <0.0001), indicating sensitivity to drought. In 252 contrast, the change in A max from week one to two was not significantly different between the 253 drought and control treatment for feral (t (132) = 0.18, p = 0.855) and domesticated (t (133) = 1.21, p = 254 0.229) plants. 255 The difference in A max between weeks 2-3 (shown as the final slope in Fig. 2A) for drought treated 256 plants indicates the degree of drought recovery (resilience). Drought treatment plants of all groups 257 showed increased A max between weeks 2-3 (positive slopes), indicating drought resilience. This trend 258 differed by domestication group (Week × Treatment × Domestication interaction, χ² = 16.06, p < 259 0.001). Wild accessions fully recovered from drought, with no significant difference in A max between 260 drought and control plants in week 3 (t (28.1) = 0.69, p = 0.497). In contrast, domesticated t (30.4) = 2.12, 261 p = 0.043) and feral (t (38.2) = 3.07, p = 0.004) accessions still show reduced A max after rewatering. 262 Wild plants in the drought treatment had the greatest increase in A max between weeks 2-3, which was 263 significantly steeper than plants in the control treatment (t (115) = -7.27, p < 0.0001), indicating rapid [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 264 recovery of assimilation rate upon rewatering. Thus, the wild varieties were both the most sensitive 265 and the most resilient to drought. 266 Pre-dawn measurements of the quantum yield of photosystem II (YII) indicate overall leaf 267 photosynthesis health. We found no significant difference in quantum yield between drought and 268 control plants of all varieties in weeks 0, 1 and 2, (p-values all >0.05; Fig. 2B). In week 3 (following 269 re-watering), wild plants showed significantly lower quantum yield for drought compared to control 270 plants (t (120) = 5.00, p < 0.0001, Fig. 2B), while there was no significant difference in quantum yield 271 between drought and control plants for domesticated (t (120) = 1.29, p = 0.199) and feral (t (120) = 1.02, p 272 = 0.309) varieties. 273 Notably, several accessions showed comparatively high A max as well as full photosynthetic recovery 274 upon rewatering in week 3, including HRIGRU 8688 (Wild; Caucasus) and CR 2299 (Feral; 275 Colombia; Fig. 2C). These genotypes are likely both highly productive and drought resilient. 276 277 **Figure 2.** Leaf maximum photosynthesis (A) and quantum yield (B) over time. Week 0 = all plants 278 were watered; week 1-2 = drought treatment received no water while control treatment received 279 water; week 3 = all plants were watered. W indicates watering and M indicates measurements. (A) 280 Steady-state CO 2 assimilation rate at 800 PPFD in control (blue) vs. droughted (red) plants by 281 domestication status. (B) Quantum yield of photosystem II (YII) in control vs. droughted plants; 282 color scheme and experiment time series are the same as in (A). Points represent group means by 283 domestication status; bars represent standard error (SE); solid lines connect group means. (C) 284 Comparison of steady-state CO 2 assimilation rate at 800 PPFD in the control group vs. the recovery 285 group (droughted and re-watered) plants upon rewatering (week 3). The dashed line indicates a 1:1 286 linear relationship. Points indicate group means by accession (blue = domesticated, green = feral, 287 gold = wild). 288 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 289 **Water Use Efficiency and Stomatal Conductance.** 290 In week 0, before the drought treatment, wild plants had significantly lower water use efficiency 291 (WUE) than domesticated plants (t (24.6) = 2.58, p = 0.042; Fig. 3A), but the difference from feral 292 plants was not significant (t (24.0) = 2.01, p = 0.131). Wild plants also had significantly greater stomatal 293 conductance (g sw ) than both domesticated (t (24.6) = -5.36, p < 0.0001) and feral (t (23.9) = -2.72, p = 294 0.031) varieties (Fig. 3B). In week 2 (two weeks post-drought), domesticated (t (22.1) = 2.85, p = 295 0.0092) and wild plants (t (20.1) = 3.66, p = 0.0016), had lower WUE in the drought compared to the 296 control treatments, while there was no difference in WUE between drought and control for the feral 297 plants (t (31.6) = 0.20, p = 0.844). This indicates that none of the groups responded to the initiation of 298 drought by becoming more water use efficient. Within the drought treatment in week 2, feral plants 299 had significantly higher WUE than wild plants (t (63.3) = 3.04, p = 0.0095). In week 3 (post re 300 watering), all varieties in the drought treatment showed an increase in WUE (domesticated: t (5.71) = 301 1.63, p = 0.16; feral: t (6.23) = -3.38, p = 0.014; wild: t (5.64) = -2.13, p = 0.08) and decrease in g sw 302 (domesticated: t (7.34) = 1.56, p = 0.16; feral: t (8.47) = 4.28, p = 0.002; wild: t (7.01) = 1.68, p = 0.14; Fig. 303 3), with feral plants having significantly higher WUE and lower g sw in the drought compared to the 304 control treatment (Fig. 3). Within the control treatment, domesticated plants had significantly higher 305 WUE than feral plants (t (36.7) = 3.00, p = 0.013). 306 Considering the change from week to week, WUE increased significantly overall (p < 0.0001) from 307 week 0 to week 1, however, this increase occurred under both drought and control treatments (Fig. 308 3A), so this was not a drought-specific response. While there was a treatment × domestication effect 309 when averaged across weeks (p = 0.0038), the rate of increase did not differ significantly for any 310 domestication type across treatments (domesticated: t (178) = -0.417, p = 0.678; wild: t (187) = 0.40, p = 311 0.687; feral: t (188) = -1.63, p = 0.106). Between weeks 1 and 2, domesticated plants showed a 312 significant difference in slope between treatments (t (238) = 2.46, p = 0.015), while wild (t (236) = 1.55, p 313 = 0.124) and feral (t (237) = 1.05, p = 0.3) plants did not. However, the three-way interaction for this 314 pattern was not significant (p = 0.6506). The treatment × domestication effect, averaged across the 315 two weeks, was the most significant factor (p = 0.0009). Between weeks 2 and 3, drought affected 316 WUE for all domestication types (week × treatment, p < 0.0001), and the effect was significant 317 within each group tested separately (domesticated: t (144) = -3.43, p = 0.0008; feral: t (161) = -2.48, p = 318 0.0144; wild: t (140) = -4.28, p < 0.0001), with wild plants showing the largest difference in slope 319 between treatments, followed by domesticated and feral plants; however, there was no significant [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 320 three-way interaction for week × treatment × domestication type (p = 0.483), indicating a trend rather 321 than group differences. Domestication type (p = 0.0087), week × domestication type (p = 0.0015), 322 and treatment × domestication type (p = 0.0075) were significant, indicating that the different 323 domestication types differed in how WUE changed as the plants recovered from drought. 324 325 **Figure 3.** Water use efficiency and stomatal conductance timeseries. (A-B) Week 0 = all plants were 326 watered; Week 1-2 = drought treatment received no water while control treatment received water; 327 Week 3 = all plants were watered. W indicates watering and M indicates measurements. (A) Water 328 use efficiency (WUE) in control (blue) vs. drought (red) treatments by domestication status. (B) 329 Stomatal conductance to water (g sw ) in control vs. drought treatments; color scheme and experiment 330 time series are the same as in (A). Points represent group mean by domestication status; bars 331 represent standard error (SE); solid lines connect group means. 332 333 In week 0, wild plants had significantly higher baseline stomatal conductance (g sw ) than domesticated 334 (t (24.6) = -5.36, p < 0.0001) and feral (t (23.9) = -2.72, p = 0.031) plants. In week 1, there were 335 significant differences between domesticated and feral plants in the control treatment (t (58.6) = -4.19, p 336 = 0.0003). In week 2, drought significantly reduced g sw in feral (t (8.02) = 3.52, p = 0.0078) and wild 337 (t (6.08) = 5.84, p = 0.0011) plants compared to the respective accessions in the control treatment. In 338 week 3, feral plants in the drought treatment had significantly reduced g sw compared to feral plants in 339 the control treatment (t (8.47) = 4.28, p = 0.0024), while domesticated and wild plants had statistically 340 similar g sw between treatments (domesticated: t (7.34) = 1.56, p = 0.161 and wild: t (7.01) = 1.68, p = 341 0.138). [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 342 Considering the rate of change in stomatal conductance, g sw declined more rapidly in wild plants in 343 the drought treatment compared to wild plants in the control treatment from week 0 to week 1 (t (135) = 344 2.22, p = 0.028) and from week 1 to week 2 (t (137) = 4.21, p < 0.0001). Between weeks 2 and 3, 345 drought reduced g sw for wild plants (t (164) = -3.45, p = 0.0007), while domesticated and feral plants 346 showed no significant differences in slope between drought and control treatments (domesticated: 347 t (168) = 0.118, p = 0.907 and feral: t (184) = 0.785, p = 0.433). 348 **Taproot Biomass and Water Use Efficiency.** Mean water use efficiency (WUE = A max g sw [-1] ), 349 calculated by accession and treatment over the 4-week time series, was positively associated with 350 taproot dry biomass (R [2] = 0.23, F(1,48) = 14.21, p = 0.00045; Fig. 4). Thus, plants with larger 351 taproots were generally more water efficient. 352 353 **Figure 4** . _B. rapa_ taproot dry biomass association with water use efficiency. (A) Points represent 354 group means by accession and treatment (control = light blue, light green, light gold; drought = dark 355 blue, dark green, dark gold), colored by domestication status (domesticated = blue, feral = green, 356 wild = gold). The solid line fits a linear regression and the gray shaded area indicates a 95% 357 confidence interval band. Taproot dry biomass data were log transformed to normalize data 358 distribution prior to regression modeling. (B) Representative variation in taproot biomass across 359 domesticated (Purple Top, CR 1477), feral (PI 311711), and wild (CR 2354, BRA 2809) accessions. 360 **Discussion** [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 361 Taking advantage of genetic variation in cultivated, feral, and wild _B. rapa_ accessions, our results 362 provide several insights into plant drought response across domestication status. We found there was 363 substantial variation in drought resistance and resilience, with wild varieties often highly resilient in 364 multiple dimensions of drought response (e.g., biomass retention, recovery of maximum CO 2 365 assimilation rate) compared to crop and feral varieties, suggesting that conspecific wild relatives of 366 crops can serve as valuable breeding material to face drought-related challenges. Specifically, we 367 found that several accessions of _B. rapa_ accumulated similar or even increased biomass under 368 drought compared to well-watered conditions. Accessions with increased biomass in the drought 369 treatment included the wild accessions CR 2354, CR 2242, and CR 2241 from the country Georgia. 370 The fact that these accessions had greater biomass under drought than watered conditions, indicating 371 overcompensation, is striking. Overcompensation to drought has been observed in multiple species, 372 but our understanding of how this occurs is limited (Ke et al., 2025). In our study, we found the 373 overcompensating accessions had slightly higher A max and g sw than the average accession; however, 374 these differences were not statistically significant. Considering the adaptive landscape of these 375 accessions, Georgia, and the Caucasus more broadly, is among the best documented regions for the 376 presence of wild _B. rapa,_ with populations there having split from domesticated forms an estimated 377 3,430–5,930 years before present (Sinskaja, 1969; McAlvay et al., 2021). This region has a highly 378 diverse climate, with patterns of precipitation and drought that have varied widely over time 379 (Elizbarashvili et al., 2025), which may facilitate the adaptation of wild _B. rapa_ populations there to 380 drought conditions. 381 There are several potential reasons that wild accessions had greater drought recovery than 382 domesticated or feral forms. One important factor is that wild accessions in the drought treatment 383 were shown to fully recover maximum CO 2 assimilation rates (compared to the control treatment) 384 after rewatering, with the combination of elevated A max and recovery of A max in wild accessions such 385 as HRIGRU 8688 (Caucasus). This indicates higher resilience to multi-week drought stress in wild 386 accessions than in feral and domesticated accessions, even with a significant decline in the quantum 387 yield (YII) of wild plants after rewatering. Since YII is an indication of photosynthetic health and has 388 been linked to xylem hydraulic failure (Brodribb et al., 2021), this suggests physiological tolerance 389 for hydraulic failure as a mechanism for fast growth, potentially as part of a drought escape strategy. 390 These results, together with reduced water use efficiency (WUE) in drought wild accessions, are 391 consistent with drought escape through inefficient water use, rapid development, and early flowering, 392 as had been previously reported in _B. rapa_ (Franks, 2011). [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 393 While the drought escape strategy is usually viewed as completing the lifecycle before the onset of 394 drought (Kooyers, 2015), our results suggest that this same combination of traits can also allow high 395 drought resilience. Thus, drought occurring early in the growing season, or brief periods of drought, 396 could both select for traits linked to rapid development that would allow plants to escape or quickly 397 recover from drought. The idea that wild varieties use water prolifically to allow for rapid growth 398 would also explain our finding that these varieties were both the most sensitive and the most resilient 399 to drought. 400 A comparison of wild and crop varieties in our drought response experiment indicates that 401 domestication diminished drought recovery in _B. rapa_, likely because crops were irrigated or grown 402 in locations or during times with abundant precipitation. This finding is consistent with the idea that 403 crops are often selected for increased yield in benign conditions of sufficient soil moisture and 404 nutrients but lack stress tolerance compared to wild varieties. which are thus a good source of genes 405 for increasing the ability of plants to respond to adverse conditions. It is important to note that 406 because of trade-offs, increasing drought tolerance in crops could potentially decrease yield or other 407 desirable crop traits (Renzi et al., 2022), so it is necessary to investigate the full effects of such stress 408 response strategies. However, we identified several varieties that had both high drought resilience 409 and high A max, indicating both high stress recovery and high productivity, likely due to rapid 410 recovery of assimilation rate upon rewatering in combination with a fast growth strategy. 411 Increased drought resilience in wild relatives has been documented in other crop species such as 412 cowpea ( _Vigna unguiculata_ ) (De Meyer et al., 2026) and several cereal species (Toulotte et al., 413 2022). In one study, domesticated sunflower ( _Helianthus annuus_ ) accessions experienced a 24% 414 decline in biomass following drought, while the wild relatives did not experience statistically 415 significant changes (Tran et al., 2024). Wild relatives of wheat vary widely in a range of drought 416 related traits, and different species and populations appear to use distinct strategies to conserve water, 417 for example through reduced stomatal conductance or differences in root architecture (Merchuk 418 Ovnat et al., 2016; Pour-Aboughadareh et al., 2017). Crop wild relatives typically retain higher levels 419 of genetic diversity than domesticated forms, and can preserve alleles that can confer resistance to 420 drought, disease, and other challenges (Dempewolf et al., 2014; Brozynska et al., 2016; Zhang et al., 421 2017). This is often due to the domestication bottleneck effect, where diversity is constrained by 422 generations of inbreeding, genetic drift, and selection, though the degree of these effects varies across 423 crops (Von Wettberg et al., 2018; Tobón-Niedfeldt et al., 2022). [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 424 Because biomass allocation and growth strategies differ across crop types (leafy, oilseed, and root 425 hypocotyl), these types could differ substantially in drought response (Farooq et al., 2019). However, 426 we did not find clear differences in drought response among the _Brassica_ crop types assayed, 427 possibly due to low replication for these groups (three leafy, two oilseed, and three turnip 428 accessions). However, this does not preclude the possibility that drought response may depend on the 429 harvested organ of the plant and may be apparent under different environmental conditions, so further 430 studies on this topic would be useful. 431 Another key finding of our study is that plants with greater taproot dry biomass had greater 432 instantaneous water use efficiency. This result is notable because it was previously thought that a 433 large taproot could help plants tolerate or recover from drought by providing resources from this 434 storage organ that plants could draw on when needed (Kalra et al., 2024), but this would not 435 necessarily increase the efficiency of water use. In contrast, we found that taproot biomass was 436 associated with greater WUE under drought conditions than plants with smaller taproots. This result 437 could help to explain why crops with taproots were often able to quickly recover after drought, as 438 was found in a recent meta-analysis (Dao et al., 2025), and shows the potential role of fleshy taproots 439 in providing a hydraulic buffer for droughted plants and promoting water use efficiency. 440 Overall, the results of this study also contribute to an emerging theory of how domestication and 441 ferality shape plant functional traits (Milla et al., 2015) and support the idea that feralization is not 442 simply domestication in reverse (Niego and Benítez-Burraco, 2022). The contrast in drought 443 response between the wild and feral accessions challenges assumptions that wild and feral accessions 444 possess similar traits and can be treated together as uncultivated germplasm. Our findings suggest 445 that these two groups diverge in drought response. These differences may be due to the different 446 adaptive environments faced by wild and feral plants, with wild plants shaped by the need for 447 tolerance to climate stress and feral plants adapting to disturbance and anthropogenic environments 448 that may be more water and fertility-rich, such as farmland edges and suburban areas (Van Alfen, 449 2014; Mabry et al., 2023). In feral plants, these factors could lead to less conservative strategies 450 around water use or less selective pressure on drought-resilient physiology (Mabry et al., 2023). 451 These findings suggest that feral populations and wild populations can contain suites of physiological 452 traits that are very different from each other, affecting their utility for breeding. In addition, feral 453 populations may also be adapted to other disturbance-heavy anthropogenic habitats such as 454 roadsides, field margins, and waste areas. Further complicating generalizations of wild and feral [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 455 populations is the fact that feral populations may arise with or without introgression with wild 456 relatives, and feral _B. rapa_ populations appear to have some proportion of ancestry from wild 457 populations (McAlvay et al., 2021). This may partially explain some of the variation observed, such 458 as the feral PI 312124 from Guatemala which had the 5 [th] highest biomass retention of the 25 459 accessions. Understanding these differences in physiological domestication and feralization 460 syndromes may help guide pre-breeding efforts to identify resilience traits in non-domesticated 461 populations (Lei et al., 2024). 462 Given the increasing intensity and duration of droughts in many areas due to climate change (IPCC, 463 2023), it is critical to understand how plants respond to and recover from drought (Xu et al., 2010; 464 Fang and Xiong, 2015). Our findings demonstrate that evolutionary history and accession identity 465 were important factors in determining drought response in _B. rapa_ . The wild relatives CR 2354, CR 466 2242, and CR 2241 represent potential candidates for further investigation as a source of useful 467 genetic variation due to their high biomass retention and rapid restoration of photosynthetic function 468 following drought. Wild and feral accessions exhibited differentiation in terms of drought response, 469 indicating that feralization does not necessarily imply a return to wild phenotypes. Wild and feral _B._ 470 _rapa_ relatives represent distinct genetic resources, and this evolutionary history should be considered 471 in breeding programs to increase stress resistance and resilience. 472 **Supplementary data** 473 Supplementary Figure S1. Experimental design. 474 Supplementary Figure S2. Soil moisture content (%) time series, by cultivar group. 475 Supplementary Figure S3. Biomass retention and crop type. 476 Supplementary Table S1. Trait table for 25 accessions of _Brassica rapa_ (separate file). 477 Supplementary Table S2. Watering scheme for experimental control and drought treatments. 478 Supplementary Table S3. Estimated marginal means (± SE) for photosynthetic assimilation rate 479 (A max ) across weeks. 480 Supplementary Table S4. Estimated marginal means (± SE) for Quantum Yield (YII) across weeks. [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 481 Supplementary Table S5. Estimated marginal means (± SE) for Water Use Efficiency (WUE) across 482 weeks. 483 Supplementary Table S6. Estimated marginal means (± SE) for Stomatal Conductance (g sw ) across 484 weeks 485 486 **Acknowledgments.** We thank the horticultural staff of the Nolen Glasshouse (NYBG) for assistance 487 with experiment setup and maintenance, and volunteer Khue Nguyen for assistance with cleaning 488 root samples. 489 **Author Contributions** 490 M.D., A.C.M., and C.T. conceptualized the study. A.C.M., A.M.B., K.I., A.E., and C.T. designed and 491 executed the experiments, with data collection and analysis contributions from G.I.Q., G.A., R.W., 492 L.R.V., M.L., R.M., I.G., R.S., J.B., L.C.D., E.M., G.M. and M.M. The manuscript was drafted by 493 A.M.B., G.I.Q., S.J.F., and A.C.M. with significant contributions from G.M. and M.M. All authors 494 contributed to manuscript review and revision. 495 **Conflict of Interest.** The authors declare no conflict of interest. 496 **Funding.** E.M., R.S., and J.B. were supported by the Science & Cultural Institutions Network (SCI 497 Network) NYC. G.I.Q. was supported by the US National Science Foundation (Franks: IOS 498 2204374). 499 **Data availability.** All data are available in the manuscript or deposited at https://osf.io/6z37u. 500 **References** 501 **Bates D, Mächler M, Bolker B, Walker S** (2015) Fitting Linear Mixed-Effects Models Using lme4. 502 Journal of Statistical Software **67** : 1–48 503 **Begna T, Wakweya RB** (2025) Climate-Smart Agriculture: Effect of Climate Change on Food 504 Security and Its Mitigation Strategies. International Journal of Agronomy **2025** : 9972955 505 **Brodribb T, Brodersen CR, Carriqui M, Tonet V, Rodriguez Dominguez C, McAdam S** (2021) 506 Linking xylem network failure with leaf tissue death. New Phytologist **232** : 68–79 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 507 **Brozynska M, Furtado A, Henry RJ** (2016) Genomics of crop wild relatives: expanding the gene 508 pool for crop improvement. Plant Biotechnology Journal **14** : 1070–1085 509 **Chikkaputtaiah C, Debbarma J, Baruah I, Havlickova L, Deka Boruah HP, Curn V** (2017) 510 Molecular genetics and functional genomics of abiotic stress-responsive genes in oilseed rape 511 (Brassica napus L.): a review of recent advances and future. Plant Biotechnol Rep **11** : 365– 512 384 513 **Dao J, Xing Y, Chen C, Chen M, Wang Z, Chen Y** (2025) Changes in shoot and root adaptations 514 of fibrous-root and taproot crops in response to different drought types: A meta-analysis. 515 Agricultural Water Management **309** : 109320 516 **De Meyer E, Van Cauter F, Vandelook F, De La Pena E** (2026) Comparison of multi-stress 517 resilience in wild and domesticated Cowpea. Scientific Reports **16** : 5109 518 **Dempewolf H, Eastwood RJ, Guarino L, Khoury CK, Müller JV, Toll J** (2014) Adapting 519 Agriculture to Climate Change: A Global Initiative to Collect, Conserve, and Use Crop Wild 520 Relatives. Agroecology and Sustainable Food Systems **38** : 369–377 521 **Elizbarashvili M, Beglarashvili N, Pipia M, Elizbarashvili E, Chikhradze N** (2025) Seasonal 522 temperature and precipitation patterns in Caucasus landscapes. Atmosphere **16** : 889 523 **Ellstrand NC, Heredia SM, Leak‐Garcia JA, Heraty JM, Burger JC, Yao L, Nohzadeh‐** 524 **Malakshah S, Ridley CE** (2010) Crops gone wild: evolution of weeds and invasives from 525 domesticated ancestors. Evolutionary Applications **3** : 494–504 526 **Fang Y, Xiong L** (2015) General mechanisms of drought response and their application in drought 527 resistance improvement in plants. Cellular and molecular life sciences **72** : 673–689 528 **Farooq M, Hussain M, Ul-Allah S, Siddique KHM** (2019) Physiological and agronomic 529 approaches for improving water-use efficiency in crop plants. Agricultural Water 530 Management **219** : 95–108 531 **Fernie AR, Yan J** (2019) De novo domestication: an alternative route toward new crops for the 532 future. Molecular plant **12** : 615–631 533 **Franks SJ** (2011) Plasticity and evolution in drought avoidance and escape in the annual plant 534 Brassica rapa. New Phytologist **190** : 249–257 535 **Franks SJ, Kane NC, O’Hara NB, Tittes S, Rest JS** (2016) Rapid genome‐wide evolution in 536 _Brassica rapa_ populations following drought revealed by sequencing of ancestral and 537 descendant gene pools. Molecular Ecology **25** : 3622–3631 538 **Gressel J** (2005) Crop ferality and volunteerism. CRC Press 539 **Hamann E, Weis AE, Franks SJ** (2018) Two decades of evolutionary changes in Brassica rapa in 540 response to fluctuations in precipitation and severe drought. Evolution **72** : 2682–2696 541 **IPCC** (2023) Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III 542 to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 543 **Jovovic Z, Andjelkovic V, Przulj N, Mandic D** (2020) Untapped Genetic Diversity of Wild 544 Relatives for Crop Improvement. _In_ RK Salgotra, SM Zargar, eds, Rediscovery of Genetic 545 and Genomic Resources for Future Food Security. Springer Singapore, Singapore, pp 25–65 546 **Kalra A, Goel S, Elias AA** (2024) Understanding role of roots in plant response to drought: Way 547 forward to climate-resilient crops. The Plant Genome **17** : e20395 548 **Ke X, Yao J, Jiang Z, Gu X, Xu P** (2025) Recover and surpass: The mechanisms of plants 549 transition upon rehydration from drought. Plant Stress **15** : 100782 550 **Khadka K, Earl HJ, Raizada MN, Navabi A** (2020) A physio-morphological trait-based approach 551 for breeding drought tolerant wheat. Frontiers in plant science **11** : 715 552 **Kogan F, Guo W, Yang W** (2019) Drought and food security prediction from NOAA new 553 generation of operational satellites. Geomatics, Natural Hazards and Risk **10** : 651–666 554 **Kooyers NJ** (2015) The evolution of drought escape and avoidance in natural herbaceous 555 populations. Plant Science **234** : 155–162 556 **Lei Z, Li Z, Zhang W, He D, Zhang Y** (2024) From wild to cultivated crops: general shift in 557 morphological and physiological traits for yield enhancement following domestication. Crop 558 and Environment **3** : 138–146 559 **Lenth R, Piaskowski J** (2025) emmeans: Estimated Marginal Means, aka Least-Squares Means. 560 **Mabry ME, Bagavathiannan MV, Bullock JM, Wang H, Caicedo AL, Dabney CJ, Drummond** 561 **EBM, Frawley E, Gressel J, Husband BC, et al** (2023) Building a feral future: Open 562 questions in crop ferality. Plants People Planet **5** : 635–649 563 **Mabry ME, Rowan TN, Pires JC, Decker JE** (2021) Feralization: confronting the complexity of 564 domestication and evolution. Trends in Genetics **37** : 302–305 565 **McAlvay AC, Bird K, Poulsen G, Pires JC, Emshwiller E** (2017) Barriers and prospects for wild 566 crop relative research in Brassica rapa. VII International Symposium on Brassicas 1202. pp 567 165–177 568 **McAlvay AC, Ragsdale AP, Mabry ME, Qi X, Bird KA, Velasco P, An H, Pires JC, Emshwiller** 569 **E** (2021) Brassica rapa domestication: untangling wild and feral forms and convergence of 570 crop morphotypes. Molecular biology and evolution **38** : 3358–3372 571 **Merchuk-Ovnat L, Barak V, Fahima T, Ordon F, Lidzbarsky GA, Krugman T, Saranga Y** 572 (2016) Ancestral QTL alleles from wild emmer wheat improve drought resistance and 573 productivity in modern wheat cultivars. Frontiers in plant science **7** : 452 574 **Meyer RS, DuVal AE, Jensen HR** (2012) Patterns and processes in crop domestication: an 575 historical review and quantitative analysis of 203 global food crops. New Phytologist **196** : 576 29–48 577 **Milla R, Osborne CP, Turcotte MM, Violle C** (2015) Plant domestication through an ecological 578 lens. Trends in ecology & evolution **30** : 463–469 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 579 **Mir RR, Zaman-Allah M, Sreenivasulu N, Trethowan R, Varshney RK** (2012) Integrated 580 genomics, physiology and breeding approaches for improving drought tolerance in crops. 581 Theor Appl Genet **125** : 625–645 582 **Nagaharu UJJJB, Nagaharu NJJJB** (1935) Genome analysis in Brassica with special reference to 583 the experimental formation of B. napus and peculiar mode of fertilization. Japanese Journal of 584 Botany 389–452 585 **Niego A, Benítez-Burraco A** (2022) Are feralization and domestication truly mirror processes? 586 Ethology Ecology & Evolution **34** : 557–590 587 **Pisias MT, Bakala HS, McAlvay AC, Mabry ME, Birchler JA, Yang B, Pires JC** (2022) 588 Prospects of feral crop de novo redomestication. Plant and Cell Physiology **63** : 1641–1653 589 **Pour-Aboughadareh A, Ahmadi J, Mehrabi AA, Etminan A, Moghaddam M, Siddique KH** 590 (2017) Physiological responses to drought stress in wild relatives of wheat: implications for 591 wheat improvement. Acta Physiologiae Plantarum **39** : 106 592 **R Core Team** (2025) R: A Language and Environment for Statistical Computing. 593 **Rendón-Anaya M, Herrera-Estrella A** (2018) The advantage of parallel selection of domestication 594 genes to accelerate crop improvement. Genome Biol **19** : 147 595 **Renzi JP, Coyne CJ, Berger J, von Wettberg E, Nelson M, Ureta S, Hernández F, Smỳkal P,** 596 **Brus J** (2022) How could the use of crop wild relatives in breeding increase the adaptation of 597 crops to marginal environments? Frontiers in Plant Science **13** : 886162 598 **Sinskaja EN** (1969) Historical geography of the cultivated flora (at the dawn of agriculture. 599 **Tobón-Niedfeldt W, Mastretta-Yanes A, Urquiza-Haas T, Goettsch B, Cuervo-Robayo AP,** 600 **Urquiza-Haas E, Orjuela-R MA, Acevedo Gasman F, Oliveros-Galindo O, Burgeff C** 601 (2022) Incorporating evolutionary and threat processes into crop wild relatives conservation. 602 Nature Communications **13** : 6254 603 **Toulotte JM, Pantazopoulou CK, Sanclemente MA, Voesenek LACJ, Sasidharan R** (2022) 604 Water stress resilient cereal crops: Lessons from wild relatives. JIPB **64** : 412–430 605 **Tran VH, Nolting KM, Donovan LA, Temme AA** (2024) Cultivated sunflower ( _H_ _ELIANTHUS_ 606 _ANNUUS_ L.) has lower tolerance of moderate drought stress than its con‐specific wild relative, 607 but the underlying traits remain elusive. Plant Direct **8** : e581 608 **Trenberth KE, Dai A, Van Der Schrier G, Jones PD, Barichivich J, Briffa KR, Sheffield J** 609 (2014) Global warming and changes in drought. Nature Climate Change **4** : 17–22 610 **Van Alfen NK** (2014) Encyclopedia of agriculture and food systems. Elsevier 611 **Von Wettberg EJ, Chang PL, Başdemir F, Carrasquila-Garcia N, Korbu LB, Moenga SM,** 612 **Bedada G, Greenlon A, Moriuchi KS, Singh V** (2018) Ecology and genomics of an 613 important crop wild relative as a prelude to agricultural innovation. Nature communications 614 **9** : 649 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 615 **Xu Z, Zhou G, Shimizu H** (2010) Plant responses to drought and rewatering. Plant Signaling & 616 Behavior **5** : 649–654 617 **Zhang H, Mittal N, Leamy LJ, Barazani O, Song B** (2017) Back into the wild—Apply untapped 618 genetic diversity of wild relatives for crop improvement. Evolutionary Applications **10** : 5–24 619 620 **Figure Legends** 621 **Figure 1.** Biomass retention (drought biomass/control biomass) comparisons across 25 accessions of 622 domesticated, feral, and wild _Brassica rapa_ . (A) Ranking of accessions by biomass retention; dashed 623 line indicates a biomass retention value of one. (B) Comparison of biomass retention by 624 domestication status. Box upper and lower boundaries represent the 75 [th] and 25 [th] percentiles, 625 respectively, and lines across boxes represent the group median. Whiskers extend from the upper and 626 lower quartiles to the group maximum and minimum, respectively. Group means not sharing a letter 627 are significantly different by pairwise comparisons with Šidák adjustment (p < 0.05). Colors indicate 628 domestication status as shown in (A). 629 **Alt text:** Ranking of biomass retention by domestication status, showing higher biomass retention for 630 wild accessions vs. feral and domesticated accessions. 631 632 **Figure 2.** Leaf maximum photosynthesis (A) and quantum yield (B) over time. Week 0 = all plants 633 were watered; week 1-2 = drought treatment received no water while control treatment received 634 water; week 3 = all plants were watered. W indicates watering and M indicates measurements. (A) 635 Steady-state CO 2 assimilation rate at 800 PPFD in control (blue) vs. droughted (red) plants by 636 domestication status. (B) Quantum yield of photosystem II (YII) in control vs. droughted plants; 637 color scheme and experiment time series are the same as in (A). Points represent group means by 638 domestication status; bars represent standard error (SE); solid lines connect group means. (C) 639 Comparison of steady-state CO 2 assimilation rate at 800 PPFD in the control group vs. the recovery 640 group (droughted and re-watered) plants upon rewatering (week 3). The dashed line indicates a 1:1 641 linear relationship. Points indicate group means by accession (blue = domesticated, green = feral, 642 gold = wild). [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.06.764879; this version posted November 11, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.06.764879) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 643 **Alt text:** A time series showing the decline and recovery of leaf photosynthetic rates and quantum 644 yield values over the course of the experiment, by domestication status. An additional panel indicates 645 the individual accessions with highest recovery of photosynthetic rate after rewatering. 646 647 **Figure 3.** Water use efficiency and stomatal conductance timeseries. (A-B) Week 0 = all plants were 648 watered; Week 1-2 = drought treatment received no water while control treatment received water; 649 Week 3 = all plants were watered. W indicates watering and M indicates measurements. (A) Water 650 use efficiency (WUE) in control (blue) vs. drought (red) treatments by domestication status. (B) 651 Stomatal conductance to water (g sw ) in control vs. drought treatments; color scheme and experiment 652 time series are the same as in (A). Points represent group mean by domestication status; bars 653 represent standard error (SE); solid lines connect group means. 654 **Alt text:** A time series showing the decline and recovery of leaf water use efficiency and stomatal 655 conductance values over the course of the experiment, by domestication status. 656 657 **Figure 4** . _B. rapa_ taproot dry biomass association with water use efficiency. (A) Points represent 658 group means by accession and treatment (control = light blue, light green, light gold; drought = dark 659 blue, dark green, dark gold), colored by domestication status (domesticated = blue, feral = green, 660 wild = gold). The solid line fits a linear regression and the gray shaded area indicates a 95% 661 confidence interval band. Taproot dry biomass data were log transformed to normalize data 662 distribution prior to regression modeling. (B) Representative variation in taproot biomass across 663 domesticated (Purple Top, CR 1477), feral (PI 311711), and wild (CR 2354, BRA 2809) accessions. 664 **Alt text:** Image of representative taproots from wild, feral, and domesticated accessions, as well as a 665 scatter plot indicating a positive linear relationship between taproot dry biomass and water use 666 efficiency.