In the winter of 2022, at the Walnut Street Greenhouse, Madison, Wisconsin (43.076222571138636, -89.42402327492367), two exPVP (PHZ51, PHP02), one heirloom (Hickory King), and three landraces (Oaxa233, Oaxa524, Oaxa733) were planted in a randomized design with six replications. Planting was done in the second week of February 2022. All landraces were obtained from the International Maize and Wheat Improvement Center (Spanish acronym CIMMYT), previously collected from southern Mexico, and the exPVP and giant heirloom were acquired from the American Germplasm Resources Information Network (GRIN). All material used in greenhouse and field experiments is listed in Supplementary Tables 1 , 2 . Ten seeds per genotype were sown in individual segments of a 24-segment plastic tray filled with HP+ Promix potting media. At 14 days after seeding, 6 seedlings of similar size per genotype were each transplanted to seven-gallon pots (Classic 2800) filled with HP+ Promix potting media. Germination rate was similar across the genotypes except for PHP02 that was 50% rate. The greenhouse temperature conditions were established as 25°C during the day and 20°C at night, with a 12-hour light cycle maintained from 6 am to 6 pm. Plants were irrigated using an automated drip irrigation system three days per week, and they were watered manually once per week with 500 ml of fertilizer water (NPK: 20-10-20). Last leaf epicuticular wax production was used as a morphological marker to determine the transition between the juvenile and adult vegetative phases (Freeling and Lane, 1994; Bongard-Pierce et al., 1996). This trait was scored on each plant at the V10-V12 stage following the protocol described by Foerster (1913); days to anthesis were summed from the planting date to 50% anther dehiscence; the number of nodes with aerial roots longer than 1 centimeter was counted on each plant at anthesis (Foerster, 1913).

In the summer of 2021, at the West Madison Agricultural Research Station, Wisconsin (43.0610329663, -89.5325872441), three exPVP (PHZ51, PHP02, HB8229) and three landrace accessions (Oaxa139, Oaxa524, GRIN19897) were planted in single-row plots in a randomized complete block design with three biological replicates. Planting was done the first week of June. These landraces were selected from previous studies and field observations in nitrogen fixation in maize (Van Deynze et al., 2018). The exPVPs were chosen as materials adapted to the Midwest in the US (Mikel, 2006; Van Deynze et al., 2018). Stand counts were conducted in each plot to assess germination rate and synchrony, and no issues with germination were observed. The last leaf with epicuticular wax was scored on three plants per plot at the V10-V12 stage as previously described; days to anthesis were summed from planting date to anther dehiscence in 50% of tassels in a plot; the number of nodes with aerial roots was counted on three plants per plot at anthesis. Additionally, the number of nodes with aerial roots was recorded weekly on three plants per plot over nine weeks, from 63 DAP to 119 DAP. In the summer of 2022 at the Hancock Agricultural Research Station in central Wisconsin (44.11985775008647, -89.53536890800639), three exPVPs (PHZ51, PHP02, and HB8229) and four landraces (Oaxa139, Oaxa524, Oaxa733, and GRIN19897) were planted in single-row plots in a randomized complete block design with three plot biological replicates and per replicate tree independent plant were measured. The number of nodes with aerial roots was recorded at 71, 101, and 116 DAP. Supplementary Table 2 provides access to information about our experiments.

In the summer of 2022 at the University of Georgia’s Iron Horse Farm, Georgia (33.72707583516916, -83.30204915982816), 11 landrace accessions (Oaxa233, Oaxa229, Oaxa306, Oaxa310, Oaxa141, Oaxa139, Oaxa524, Oaxa612, GRIN19970, Oaxa622, Oaxa182) were planted and screened in single-row plots organized by genotype in a complete block design with three replicates. Planting was completed in the second week of May. Stand counts were conducted in each plot to assess germination rate and synchrony, and no issues with germination were observed. The phenotypes evaluated during the Wisconsin field study 2021 were also assessed, although the aerial root phenotypes were checked on five plants per plot. Supplementary Table 2 provides access to information about our experiments.

Three maize accessions (GRIN19770, Oaxa139, and Oaxa229) were planted at the University of Wisconsin-Madison’s Walnut greenhouse (Madison, WI) under controlled conditions of 25°C during the day, 20°C at night, and a 12-hour light cycle from 6 am to 6 pm. Due to the light-sensitive nature of these maize accessions, all windows and walls were covered with black tarps. Seven-gallon pots (Classic 2800) were filled with a mixture of autoclaved calcined clay, also known as Turface(R) (Turface Athletics MVP), and sand in a 4:1 v/v ratio, adapted from Kottkamp et al. (Kottkamp et al., 2010). The high temperatures during calcination cause the clay to expand, creating a porous structure. This structure provides physical and chemical stability, allowing for aeration to plant roots while retaining water within its pore network. This substrate selection aimed to minimize variability in soil potting media, ensuring consistent growing conditions for the plants. Each pot was positioned over a 13-inch plastic pot saucer to prevent the loss of fertilization treatment through the pots’ drainage holes. These trays served as a means to ensure that all the plants had access to the same amount of nitrogen treatment. To investigate the effect of nitrogen fertilization on the number of nodes with aerial roots, three different nitrogen rates were tested: 175 ppm (high), 120 ppm (medium), and 70 ppm (low). These were achieved using various ratios of Hoagland’s solutions with and without nitrogen (HOP01 and HOP03 from Caisson Labs). The nitrogen in the HOP01 mix was provided by ammonium phosphate, calcium nitrate, and potassium nitrate, which together contributed 175 ppm of nitrogen. The HOP03 mix did not contain any source of nitrogen. To achieve the desired nitrogen level, a plastic container was filled with a mix of 70% HOP01 solution and 30% HOP03 solution. For the low nitrogen level, 40% of the HOP01 solution and 60% of the HOP03 solution were used. Each treatment comprised three replicates, resulting in a total of 27 pots that were randomly distributed across the greenhouse room. Synchronized germination was similar across the genotypes. Plants were watered five days a week, receiving 500 mL of the corresponding nitrogen treatment twice a week. The number of nodes with aerial roots was assessed thrice during the experiment at 80, 90, and 132 DAP. Aerial roots were counted as such only if they were not touching the ground and measured 1 cm or more long. In contrast, the root diameter of the aerial roots was measured once at 132 DAP, and three roots were selected randomly from the fourth node (or the highest node if no roots were present in the fourth node). Supplementary Table 2 provides easy access to information about our experiments.

The maize genotypes Oaxa524 and PHP02 were cultivated in two rooms at the Walnut Street Greenhouses at the University of Wisconsin-Madison. They were grown in seven-gallon pots (Classic 2800) filled with Pro-Mix LP15 medium, with daytime temperatures maintained at 28°C and nighttime temperatures at 25°C. Seeds were scored for germination with no deviation between genotypes. Pots were randomized in the greenhouse, and plants were exposed to a 12-hour light photoperiod from 6 am to 6 pm and 12 hours of darkness. A humidifier system (Smart Fog, Reno, Nevada) was employed to regulate humidity levels, ensuring a high humidity of 75% and a low humidity of 30%. Plants were irrigated twice a week with municipal water and received 500 mL of a 20-10–20 NPK fertilizer solution weekly. At the flowering stage, the plants were phenotyped for three parameters: the number of nodes with aerial roots, the number of roots on the top node, and aerial root diameter with a caliper. These measurements were carried out using the established procedure as previously described. Supplementary Table 2 provides easy access to information about our experiments.

Statistical analysis and data plotting were done using R statistical software (version 4.2.1) (R Core Team and Others, 2013). For parametric data, ANOVA and pairwise comparisons with Tukey’s Honestly Significant Difference Tukey HSD were performed with the package “multcompView (version 0.1 - 8)” (Graves et al., 2024). For non-parametric data, a Wilcoxon test was performed with the package “ggpubr (version 0.6.0)” (Kassambara, 2023). Correlation analysis was performed with the package “correlation (version 0.8.6)” (Makowski et al., 2022). ANOVA assumptions were reviewed with the package “emmeans (version 1.11.1)” (Estimated Marginal Means, aka Least-Squares Means, n.d.). All plots were generated with the package “ggplot2 (version 3.5.0.9000)” (Wickham, 2009). Pairwise comparisons for all experiments are provided in Supplementary Data 1 .

A greenhouse study was conducted to examine growth phases and aerial root formation in two exPVP inbreds (PHZ51, PHP02), one heirloom (Hickory King), and three Oaxacan landrace accessions (Oaxa233, Oaxa524, Oaxa733) under controlled conditions ( Figure 1C ). Leaf epicuticular wax was used as a marker for the juvenile-to-adult vegetative phase transition ( Figures 1D, E ). Growth of the last leaf with epicuticular wax was only significantly different for the landrace Oaxa233 compared with the other genotypes ( Figure 2A ; Supplementary Data 1 ). PHZ51, PHP02, and Oaxa524 transitioned to the adult phase after the growth of leaf nine, whereas Hickory King, Oaxa233, and Oaxa733, on average, transitioned one to two leaves later. The exPVP genotypes generally exhibited narrower variability for this trait. In contrast, the heirloom and landrace genotypes showed wider-ranging values, probably because the exPVP lines are genetically uniform inbreds, while the heirloom and landraces are heterogeneous and outbred. Anthesis (male flowering) marks the transition from the adult vegetative to the adult reproductive phase. Under greenhouse conditions with a 12-hour photoperiod, days to anthesis varied significantly, with the exPVPs and Hickory King tasseling earlier than the landrace genotypes Oaxa524 and Oaxa733 ( Figure 2B ). PHZ51 and PHP02 reached anthesis in less than 70 days, Hickory King in an average of 85 days, and the landraces in over 95 days post-planting. Oaxa233 appears to be photoperiod sensitive, and only one replicate produced a mature tassel thus was removed from the analysis. At anthesis, the number of nodes with aerial roots was quantified. The exPVP and heirloom genotypes produced significantly fewer nodes with aerial roots than landrace accessions ( Figure 2C ). For example, PHZ51 produced significantly fewer nodes with aerial roots than Oaxa524 (p = 0.029) and Oaxa233 (p = 0.0004). In addition, the adult vegetative phase (the time between forming the last leaf with epicuticular wax and anthesis) was longer in the landrace accessions than in the exPVP and heirloom genotypes ( Supplementary Figure 1 ). Potential correlations were investigated among the number of nodes with aerial roots, days to anthesis, and the last waxy leaves. Our analysis revealed positive correlations among all traits in landraces and exPVP ( Supplementary Figure 2A ). Specifically, we observed a moderately positive correlation between the number of days to anthesis and the number of nodes with aerial roots (r = 0.66). In contrast, the correlations between the other pairs were very low ( Supplementary Figure 2B ). The number of nodes with aerial roots was also assessed at various intervals between 64 and 135 days after planting (DAP) ( Figure 3 ). PHP02 and PHZ51 grew no additional nodes after tasseling (average 70 DAP) and averaged three nodes with aerial roots in both exPVPs ( Figure 3 ; Supplementary Figure 1 ). In contrast, the larger Hickory King variety flowered at an average of 85 DAP and averaged five nodes with aerial roots per plant. Intriguingly, the landraces continued producing aerial roots even after anthesis, typically after 100 DAP ( Figure 3 ; Supplementary Figure 1 ). These findings suggest some Oaxacan landraces (Oaxa524 and 733) continue developing aerial roots into the adult reproductive stage, unlike the other maize lines tested here.

Insufficient nitrogen hinders plant growth and affects root architecture traits (Stitt and Krapp, 1999; Anas et al., 2020). A greenhouse experiment tested how nitrogen levels impact aerial root growth in three Oaxacan landraces. Nitrogen was applied at 70, 120, and 170 ppm. The number of nodes with aerial roots was recorded at 80, 90, and 132 DAP ( Figure 4 ). No significant differences were observed when comparing the means of each nitrogen treatment within each accession and date, except for accession Oaxa 139, which showed a significant difference between high nitrogen and the other two treatments at 80 DAP. Nitrogen levels did not affect traits such as aerial root diameter, stalk diameter, and plant height ( Supplementary Figure 7 ). These results indicate that varying levels of available soil nitrogen do not significantly impact the development of traits related to aerial root development in Oaxacan maize landraces.

The Oaxacan landraces grow in a high-humidity environment, which enables mucilage secretion from their aerial roots (Van Deynze et al., 2018; Bennett et al., 2020; Pankievicz et al., 2022). Our study aimed to investigate the effect of humidity on the development of aerial roots in the Oaxacan landrace Oaxa524, which is known for producing numerous nodes with aerial roots under various field and greenhouse conditions. We also used the exPVP PHP02 as a control. We tested two different humidity levels, 30% and 75%. Our experiment focused on measuring the number of nodes with aerial roots, their diameter, and the number of aerial roots at the top node. These parameters are associated with mucilage production (Pankievicz et al., 2022). In the Oaxa524 landrace, there was a slight increase of one node with aerial roots under high humidity (p = 0.021). However, this effect was not observed in PHP02 ( Figure 5A ). Additionally, no significant differences were observed in root diameter or the number of roots at the top node under varying humidity conditions ( Figures 5B, C ). These findings indicate a positive but limited effect of humidity in promoting the development of aerial root nodes in this Oaxacan maize landrace.

The development of above-ground nodal roots (brace and aerial roots) in maize has traditionally been associated with the juvenile stage (Poethig, 1990; Hostetler et al., 2021). Mutants such as Corngrass1 (cg1) are well known to develop five to six nodes with aerial roots due to an extended juvenile phase, as delayed flowering occurs 1 or 2 weeks, depending on the inbred background (Poethig, 1988, 2003; Chuck et al., 2007). Numerous maize accessions can develop such roots during the adult stage in response to lodging (Sparks, 2023). However, we showed here that the increased number of nodes with aerial roots in Oaxacan accessions is not due to an extended juvenile phase but to the production of these nodal roots at the adult stage without lodging. Apart from inherent diversity, numerous genes influence nodal root growth, acting as negative regulators (Hostetler et al., 2021). For example, knocking down the teosinte glume architecture1 (tga1) gene resulted in an increased number of nodal roots ( Wang et al., 2015 ). Intriguingly, many of these genes are interconnected with flowering and responses to gibberellic acid, as reviewed in (Hostetler et al., 2021). Nonetheless, flowering in maize is a complex trait governed by many loci with modest impacts (Buckler et al., 2009). This trait is a pivotal adaptation linked to geographical positioning (Ziello et al., 2009; Crimmins et al., 2010). Our observations reveal that the landraces tested here (all cultivated above 1,500 meters) exhibit delayed/late flowering compared to the exPVP lines. Understanding the genetic factors underlying these two traits requires a thorough investigation. Despite research indicating no connection between flowering time and brace root development, it is imperative to investigate the potential presence of genetically correlated variables in these landraces (Reneau et al., 2020). Tropical genotypes are characterized by late flowering, significant height, and increased biomass (leaves) when grown in temperate latitudes (Warrington and Kanemasu, 1983). Wisconsin and Georgia experience distinct photoperiods, with Wisconsin receiving up to 15 hours of daylight at its peak, in contrast to Georgia’s 14 hours. This one-hour difference in daylight may account for why landraces tend to flower earlier in Georgia than in Wisconsin, reflecting the tropical origins of this material. Hence, delving into these variables is pertinent to gain deeper insights into their suitability for breeding initiatives. We evaluated the epicuticular wax as a morphological marker to determine the transition between juvenile and adult stages. In maize, juvenile leaves have epicuticular wax layers, whereas adult leaves lose this waxy layer and have crenulated epidermal cells (Sylvester et al., 2001). The environment affected this trait, as waxy leaves in Oaxa139 and Oaxa524 genotypes ceased to occur at an earlier growth stage in GA than in WI ( Figure 2 ; Supplementary Figures 3 , 5 ). In our investigation, through scrutinizing the number of nodes with roots and various traits across three distinct environments, we observed that the landrace Oaxa524 exhibited higher performance compared to other landraces, particularly in terms of number of nodes with aerial roots ( Figures 2 , 3 ; Supplementary Figures 4 - 6 ). This insight holds promise for breeding programs that incorporate aerial root traits into varieties suited for temperate climates, while introducing nitrogen fixation capabilities to reduce reliance on fertilizers. We recognize the need for further studies to determine the adaptability of these root traits across various environments, ranging from temperate to tropical regions. The selected exPVP inbred lines served as references, due to their extensive phenotypic and genotypic characterization and adaptation to temperate environments. However, we acknowledge the limitations of this comparison, particularly regarding maize diversity, as landraces are a source of unique alleles. In the future, generating hybrids between these exPVP lines and landraces will offer a more robust framework for assessing traits under local environmental conditions and better capturing heterotic effects. Additionally, the inclusion of tropical inbred lines of Mexican origin, such as the CIMMYT Maize Lines (CMLs), which were developed using landraces from Mexico. (Warburton et al., 2008) could significantly improve these comparisons. The CMLs are genetically closer to Oaxacan landraces and exhibit root phenotypes that are more similar to those found in the Oaxacan landraces. Future studies should incorporate both exPVP-derived hybrids and a broader panel of tropical lines, including CMLs, to enhance the generality of our findings and more accurately contextualize the unique traits of Oaxacan maize. Moreover, it has been noted that the aerial root development trait is not exclusive to Oaxacan landraces but is instead distributed across other regions of Mexico (Connolly et al., 2025). This expanded framework will also help clarify whether the observed phenotypic differences are primarily due to environmental adaptation or inherent genetic divergence across breeding histories. Currently, our analysis indicates that above-ground nodal root development should not be used as a marker for the juvenile stage in maize, at least among Oaxacan-like materials.

Given the well-documented impact of nitrogen limitation on various physiological processes, many farmers worldwide rely on synthetic fertilizers to boost yields. However, their excessive use poses environmental risks, including water contamination through runoff and economic challenges, especially for small-scale farmers, who often experience the direct effects of price fluctuations influenced by global events (Arndt et al., 2023). With the growing interest in reducing synthetic nitrogen fertilizer use by enhancing nitrogen delivery through biological nitrogen fixation (BNF), it is crucial to understand whether an optimal nutritional rate supports the proper development of aerial roots, thereby maximizing the benefits of BNF. Our greenhouse experiment did not reveal any significant effect of soil nitrogen on aerial root development ( Figure 4 ). In the future, it would be interesting to assess a broader range of nitrogen levels, encompassing both lower and higher concentrations than those used in this study. Additionally, the greenhouse experiment used Turface^® as the substrate, which, while advantageous for controlling nitrogen levels, differs considerably from field soil. Therefore, conducting similar experiments under field conditions in the future would be particularly interesting.

Plants display a variety of responses to changes in humidity levels. The biomass of aerial roots in indoor plants is increased due to the high relative humidity within households, facilitated by the indoor environment and misting systems, similar to those in our study (Sheeran and Rasmussen, 2023). Humidity affects the concentration of phytohormones. For instance, cucumbers’ gibberellic acid and auxin levels decrease in response to high humidity (Amin et al., 2021). Similarly, the ability to produce more aerial roots was found to be strongly affected by high relative humidity in tomatoes, resulting in a slight reduction in abscisic acid (ABA) levels and an increase in ethylene (Arve and Torre, 2015). Notably, applying exogenous auxin to stem cuttings increased adventitious roots in tomatoes, underscoring the significant influence of hormones and how various stimuli can further enhance this response (Guan et al., 2019). In maize, adventitious roots emerge from aerial nodes, and their development, elongation, and quantity are influenced by a complex interplay of genes and hormones (Singh et al., 2023). Elevated humidity slightly enhanced the number of nodes bearing aerial roots in the landrace Oaxa524 ( Figure 5A ). A phloem-based auxin response may contribute to this increased root formation in maize, akin to what has been observed in rice under saturated humidity (Chhun et al., 2007). Furthermore, ethylene, a plant hormone that accumulates during flooding and induces adventitious nodal roots, could be pivotal in developing nodes with aerial roots in maize. 1-Aminocyclopropane 1-carboxylic acid, which serves as an ethylene precursor, stimulates the generation of adventitious nodal roots in maize (Singh et al., 2023). Consequently, hormone precursors or their inhibitors emerge as a valuable approach to assess the progression of aerial root development in cereal crops. However, humidity did not impact the diameter or the number of roots in our maize accessions ( Figures 5B, C ). In contrast, a study of above-ground roots within the Araceae family showed that the root diameter increases under high humidity conditions (Sheeran and Rasmussen, 2023). Our results suggest that signals such as phytohormones may govern aerial root elongation in Oaxacan maize landraces at the adult stage without any lodging stimulation. Moreover, our study indicated that the impact of soil nitrogen and ambient humidity on aerial root development is limited, which may facilitate the use of this trait across diverse environments. However, as our experiments were conducted in a greenhouse, field studies are needed to confirm these findings.

In this study, our main objective was to explore how specific environmental factors influence the development of aerial roots in Oaxacan maize landraces and compare them with some exPVP adapted material to the Midwest. We recognize that these variables were analyzed in isolation, underscoring the importance of integrating them in the future to replicate field conditions more accurately. Much of the research was conducted in controlled greenhouse settings, limiting the impact of other environmental factors on aerial root development. Further investigation under diverse field conditions will be necessary to validate these findings. Our long-term goal is to enhance maize’s ability to acquire nitrogen through BNF, and thereby reduce its reliance on exogenous fertilizers. The finding that Oaxacan landraces continue to produce aerial roots into the adult stage has important consequences for breeding since this trait could be introgressed into elite varieties without the potential negative impacts of an extended juvenile phase, such as delayed flowering. The minimal but statistically significant GxE interaction we observed among field sites implies some environmental component to this trait, which may make some environments (e.g., arid ones) unsuitable for trait expression or at least suboptimal. Further research on aerial root production may help identify environment-independent alleles that could be used for breeding. Identifying the specific genes involved would enable more targeted allele mining or mutagenesis to generate the desired traits. Considering the extensive scale of global maize production, even a small reduction in nitrogen requirements could have significant economic and environmental impacts worldwide. This study set the foundation for incorporating these landraces into our breeding program, aiming to develop crop varieties that require less synthetic nitrogen fertilizer and contribute to the long-term goal of sustainable agriculture. This concept has been explored by other authors, who have studied different sets of maize landraces from various regions of Mexico and proposed screening across the Americas for these genetic resources (Connolly et al., 2025). In the future, we aim to share insights with the community on the agronomic performance and yield components of these landraces.