Enhancing deep sowing success: genetic diversity in mesocotyl and coleoptile length, and field establishment of oats (Avena sativa)

Experiment 1: wide screening experiment

A wide screening experiment was conducted comparing the mesocotyl and coleoptile lengths of 195 diverse genotypes of oats (Avena sativa) (see Supplementary Table S1). The 195 genotypes were arranged using a completely randomised design with six replicates per genotype. The seeds were pre-screened to between 35 and 55 mg (husk not removed). Seeds were placed on the surface of 40 mm² Jiffy pots (http://www.jiffypot.com) filled with course vermiculite and lightly covered with a thin layer of vermiculite (5 mm depth). These conditions were selected based on protocols from those previously developed for wheat (Triticum aestivum) (Rebetzke et al. 2007) and adapted for oats through a pilot experiment, which revealed that coleoptile and mesocotyl lengths were not significantly affected by sowing depth (5 mm or 150 mm) or substrate type (potting mix or vermiculite) when grown in darkness (Table S2). Jiffy pots were organised in a 112-cell tray, with a row spacing between each block to minimise entanglement. Trays were placed in lightproof black tubs and watered (about 3 L) so that water availability did not limit seedling growth. The tub lids were taped at the corners to keep light out, as even brief exposure to light impedes cell division in the mesocotyl, resulting in growth inhibition (Mer 1966). The experiment was conducted in a room at a constant temperature of 20 (±0.7) °C for 240°-days (°Cd; 12 days). The experiment’s duration was determined based on Radford and Key’s (1993) findings that 231.4°Cd was sufficient to complete elongation of both the mesocotyl and coleoptile. After 12 days, the seedlings were carefully removed from the vermiculite, and mesocotyl (oats), coleoptile (all cereals), and the combined length for each seedling was measured by extending them along a ruler. The mesocotyl-to-coleoptile length ratio (MCR) was calculated for the oat genotypes to analyse the relationship between the mesocotyl and coleoptile length:

The broad-sense heritability of the mesocotyl, coleoptile, and combined length was calculated based on Murphy et al.’s (2008) formula:

where hB2¯ is broad-sense heritability, VG is the mean sums of squares of the genetic component of variation, and VT is the mean sums of squares of the total variation in the population

In addition to analysing the 195 oat genotypes as a whole set, they were also categorised and analysed as three distinct groups: (1) 20 Australian commercial genotypes; (2) 71 international; and (3) 104 pre-breeding lines. Australian genotypes represent those that are (or were) commercially available in Australia. International genotypes include a wide genotype of oat genotypes sourced worldwide. Pre-breeding lines were unreleased genotypes that are still under selection by the InterGrain National oat breeding program.

Experiment 2: multiple temperature screening

This experiment evaluated the variation of mesocotyl and coleoptile lengths of 42 oat genotypes across four temperatures: (1) 15 ± 0.6°C; (2) 20 ± 0.7°C; (3) 25 ±0.7°C; and (4) 30 ± 0.9°C. The experiment was set up following the same method as the wide screening experiment (see above, Experiment 1). It included 42 genotypes with six replicates completely randomised within each temperature block. They were located in four constant-temperature rooms with each set to a different temperature and running simultaneously. The 42 oat genotypes were chosen to encompass all tested in the field validation experiment (Experiment 3), plus pre-breeding genotypes that offered genetic diversity from within the InterGrain breeding program and overseas genotypes found to have to have contrasting MCR in Experiment 1 (Table S1).

For consistency, at each temperature regime, mesocotyl and coleoptile length data were collected at 240°Cd, so the 15°C, 20°C, 25°C and 30°C treatments were assessed as above at 16, 12, 10, and 8 days, respectively.

Experiment 3: field experiments

Five field experiments were conducted comparing the emergence of 25 Australian cereal genotypes: 19 oats; four wheats; and two barleys. The field experiments were conducted at four different sites in Western Australia: (1) Merredin (31°28′54.9″S, 118°13′44.5″E); (2) Kunjin (32°17′03.0″S, 117°44′01.8″E); (3) Highbury (33°04′14.0″S, 117°15′46.9″E); and (4) Kurrenkutten (32°15′58.5″S, 118°00′09.7″E). Two experiments were carried out in 2023 at Merredin and Kunjin, followed by three more in 2024 at Merredin, Highbury, and Kurrenkutten analysed as a Multi-Environment Trials (MET) experiment.

Each of the field experiment used a randomised split plot design with two replicates of each genotype, with sowing depth (shallow at 40 mm and deep at 110 mm) as the main plots and genotypes as subplots. Seeds for the field experiments were obtained from InterGrain, an Australian cereal breeding company, sourced from the 2022 commercial seed purification increase, with seed weights ranging between 35 and 55 mg (husk not removed). Seed viability and quality were confirmed before the experiment, using standard germination tests.

The oat genotypes selected were those present in the 2023 season of the National Genotype Trials and relevant to Western Australia. Four yet-to-be-released oat genotypes were also included in the experiment. Wheat and barley genotypes were selected based on varying coleoptile lengths, year of release, and significance, to the Western Australian cereal cropping system. To contrast the semi-dwarfed wheat genotypes, the historical wheat genotype Halberd was also included. A full description of each genotype is in Table 1.

Plots at all environments comprised seven rows spaced 0.25 m apart (total plot width of 1.5 m) and were either 12 m long at Merredin or 11 m long (Kunjin, Highbury, and Kurrenkutten). The number of seeds per plot for each genotype was adjusted by individual seed weight (Table 1) with a target plant density of 240 plants per m². Trials were managed by local grower practices, with adequate fertiliser and standard herbicide application, and sprayed twice with fungicide to prevent foliar disease. Two replicate emergence counts of 0.5 m² were randomly taken in each plot around 14 days after sowing (DAS), avoiding edge rows (Table 2).

Emergence counts were analysed as the ratio of emergence for deep-to-shallow sown seed:

This approach was used to eliminate the influence of variations in seed viability between genotypes. It uses emergence from shallow depth as the reference for emergence from depth. A higher deep-to-shallow emergence ratio indicates an increased ability of the genotypes to emerge from deep-sown (110 mm) compared to shallow-sown (40 mm) treatments.

Statistical analysis

The statistical analysis was conducted with the R program (R Core Team 2023).

In the first (wide screening) and second (temperature screening) experiment, the Shapiro–Wilk Test was used to check for data normality (Shapiro and Wilk 1965). Levene’s Test was used to ensure normality and error variance homogeneity between groups (Levene 1960). When the null hypothesis of equal group variances was rejected (P < 0.05), a heteroskedastic robust covariance matrix was implemented under the R statistical package ‘sandwich’ (Zeileis 2006). The distribution of the mesocotyl length, coleoptile length, combined length, and MCR in the wide screening experiment and temperature experiment (Experiments 1 and 2) were first examined with descriptive statistics from boxplot summaries.

ANOVA was then performed using the generalised linear model procedure included in the base R statistics package. The analysis examined the variation in mesocotyl length, coleoptile length, combined length, and MCR between genotypes within oats, within the three oat groups (Australian, international, and pre-breeding). Additionally, ANOVA was used to determine the effect of different temperatures on these traits. If ANOVA detected significance at P < 0.05, a pairwise post hoc test (Tukey’s ‘Honestly Significant Difference’; HSD) was implemented to test for significant differences between means. The effect of blocks was also evaluated in the ANOVA.

The third (field) experiment was analysed as a Multi-Environment Trial (MET) experiment. The variation in emergence ratio between genotypes and cereal crop species (wheat, barley, and oats) was analysed using Levene’s test, ANOVA, and pairwise post hoc test using the same methods outlined above.

Furthermore, the linear regression model, implemented under the generalised linear model procedure in the base R statistics package, was used to estimate the relationship between temperature and emergence ratio with combined length and to analyse the correlations between mesocotyl and coleoptile length.

Experiment 1: wide screening experiment

There were significant differences in combined length, mesocotyl length, coleoptile length, and MCR between the oat genotypes (P < 0.001; Table S3). Across the 195 oat genotypes, combined lengths ranged from 112 mm to 219 mm, with a mean of 169 (±2) mm (Fig. 1). The mesocotyl length was more variable and shorter than the coleoptile length, with a mean of 73 (±1) mm compared to 96 (±1) mm, respectively. The mesocotyl lengths ranged from 32 mm to 119 mm, while the coleoptile lengths ranged from 21 mm to 71 mm. Furthermore, the mesocotyl-to-coleoptile length ratio (MCR) ranged from 0.31 to 1.92, with a mean of 0.8 (±0.03). Notably, 75% of the data (3rd quartile) exhibited MCR of 0.9 (Fig. 1). Mesocotyl length and coleoptile length were not correlated (R² = 0.004).

Fig. 1.

Boxplot of total combined length (TL), mesocotyl length (ML), coleoptile length (CL), and mesocotyl-to-coleoptile ratio (MCR) of the 195 oat genotypes in the wide screening experiment, where seedlings were grown at 20°C for 240°Cd. The boxplot indicates the median (middle pinch line), interquartile range (box length – lower quartile represent the 25th percentile and the upper quartile represent the 75th percentile), and overall spread of the data (whisker – smallest and largest sample value), along with any potential outliers (indicated by black dots). Each variety means was also represented in the coloured dots.

The combined length and mesocotyl length of the 195 oat genotypes were highly heritable with hB2 of 0.81 (±0.10) and 0.85 (±0.10), respectively. This suggests that the majority of the observed (phenotypic variance) variances in oat genotypes’ combined length and mesocotyl length were attributable to genetic factors. In contrast, the coleoptile length was less heritable, with a hB2 value of 0.54 (±0.10), indicating that roughly half of the observed differences in coleoptile length can be attributed to genetic factors.

There were significant differences in combined length, mesocotyl length, and coleoptile length between the oat groups: Australian commercial; international; and pre-breeding genotypes (P < 0.001; Table S4). No significant differences were seen in the MCR (P = 0.40; Table S4). However, post hoc tests revealed no significant difference between the pre-breeding and Australian genotypes in the mesocotyl, coleoptile, or combined lengths (P = 0.07; P = 0.96; P = 0.99). International genotypes have greater combined lengths, measuring 20 mm more than Australian commercial genotypes and 15 mm more than pre-breeding genotypes (Table S4). They also have longer coleoptile and mesocotyl lengths (Fig. 2).

Fig. 2.

Four boxplots of coleoptile length (mm), mesocotyl length (mm), mesocotyl-to-coleoptile length ratio (MCR), and combined (mesocotyl and coleoptile) length (mm) of 195 oat genotypes through different oat categories (Australian, international, pre-breeding). The boxplot indicates the median (middle line), interquartile range (box length – lower quartile represent the 25th percentile and the upper quartile represent the 75th percentile), and overall spread of the data (whisker – smallest and largest sample value), along with any potential outliers (indicated by dots).

Experiment 2: temperature response

There was a strong response to temperature with increasing temperature from 15 to 30°C significantly decreased the combined length, mesocotyl length, and coleoptile length of 42 oat genotypes (P < 0.001; Fig. 3; Table S5). The mean combined length decreased from 185 (±4) mm, 165 (±4) mm, 132 (±3) mm, to 103 (±3) at 15°C, 20°C, 25°C, and 30°C, respectively. There was a significant interaction between genotype and temperature to combined length (P < 0.001), indicating that some genotypes were more sensitive to temperature than others; however, a consistent reduction in combined length with increasing temperature was observed across all the oat genotypes. The MCR increased with temperature, but this was only significant at 30°C (P < 0.001; Table S5), suggesting that coleoptile length decreased to a greater extent than mesocotyl length in higher temperatures. At 25°C and 30°C, the minimum combined length was 89 mm and 68 mm, respectively. Coleoptile and mesocotyl lengths were reduced by 3.67 mm and 1.05 mm per °C increase in temperature (Fig. 3).

Fig. 3.

Four boxplots of total combined length (mm), coleoptile length (mm), mesocotyl length (mm), and mesocotyl-to-coleoptile length ratio (MCR) of 42 oat genotypes through different temperatures (15°C, 20°C, 25°C, 30°C). The boxplot indicates the median (middle line), interquartile range (box length – lower quartile represent the 25th percentile and the upper quartile represent the 75th percentile), and overall spread of the data (whisker – smallest and largest sample value), along with any potential outliers (indicated by dots).

Field experiments

There were no significant differences between emergence ratio at an environment (P = 0.18), replicate (P = 0.88), or experiment and genotype interaction (P = 0.58) levels. There were significant differences in emergence ratio between crop species (Fig. 4; P < 0.001). Oats had a higher average emergence ratio (0.79 (±0.01)) than wheat and barley genotypes (0.56 (±0.03) and 0.51 (±0.09), respectively). However, there were no significant difference in emergence ratio between wheat and barley genotypes (P = 0.97). Most oat genotypes have a higher emergence ratio value than the wheat and barley genotypes, except for Halberd, the older non-dwarfed wheat genotype with a longer coleoptile (Fig. 4). There was no significant variation in emergence ratio across the oat genotypes (P = 0.76).

Fig. 4.

Deep-to-shallow emergence ratio of the oat, wheat, and barley genotypes 14 days after sowing (DAS) at the five environments: Merredin and Kunjin in 2023; Merredin, Kurrenkutten, and Highbury in 2024. Seedlings are deep-sown at 110 mm and shallow-sown at 40 mm. Bars are mean ± s.e., n = 4.

The mesocotyl, coleoptile, and combined length data for the genotypes included in the field trials were extracted from the wide-screen results (see Fig. 5) and showed significant differences in combined lengths (Fig. 5; P < 0.001). Compared to oat genotypes, barley and wheat genotypes exhibited lower combined length differences by 89 mm and 57 mm, respectively. All the oat genotypes displayed longer combined lengths compared to the wheat and barley genotypes, except for Halberd. Halberd, which is the older, non-dwarf wheat genotype, exhibited the longest coleoptile length at 162 (±14) mm, while other wheat and barley genotypes in the trial ranged from 58 mm to 82 mm in combined lengths (i.e. medium length class).

Fig. 5.

Combined (mesocotyl and coleoptile) length (mm) of oat, wheat, and barley genotypes included in the field experiments. Coleoptile and mesocotyl lengths were measured on seedlings grown in the dark for 20°C for 240°Cd. Oat genotypes have two stacked bars, with mesocotyl length indicated in the lower bar. Bars are mean ± s.e., n = 6.

Furthermore, there were significant differences between emergence ratio and genotype (P < 0.001), meaning that observations were consistent across all environments and thus could be analysed as replicates. The type of crop (wheat, barley, or oat) was found to have a significant effect on emergence ratio (P < 0.001). Any genotype with combined lengths greater than 110 mm (deep sowing depth in this study) had a similar emergence ratio, with no significant correlation (R² = 0.01; P = 0.57). This group included all the oat genotypes and the older wheat genotype Halberd. However, the second group, with combined lengths less than 110 mm, did not emerge well from depth, with low emergence ratio values (Fig. 6).

Fig. 6.

Relationship between the deep-to-shallow emergence ratio and combined length (mm) of the barley, oat, and wheat genotypes included in the field trial. Seedlings are deep-sown at 110 mm and shallow-sown at 40 mm.

Implications of oats in early and deep-sown conditions

Early sowing of winter cereals offers numerous advantages, including deeper root penetration, increased weed competition, greater biomass accumulation, higher water use efficiency (WUE), and reduced exposure to heat as it enables crops to flower early (Passioura and Angus 2010; Kirkegaard et al. 2015; Rasmussen and Thorup-Kristensen 2016). Early sowing could expose the seedlings to maximum soil temperatures above 30°C in the Western Australian wheatbelt during the sowing window (Sudmeyer et al. 2016). Climate change compounds this issue, as autumn weather is predicted to become hotter, drier and more variable. Some climate models indicate that by 2060, sowing season air temperatures could rise by as much as 4°C (Whetton et al. 2015; Frederiksen and Osbrough 2022). Our findings revealed that at 25°C and 30°C, the minimum mean combined length of the oat genotypes were 89 mm and 68 mm, respectively. This suggests that when sowing early with high soil temperatures, the combined length of the shortest genotypes may be below the 110 mm threshold and therefore may have limited emergence from deep sowing. Radford and Key (1993) found that higher temperatures caused temporary dormancy, resulting in further poor field emergence of oats at high soil temperatures, causing germination delays and insufficient seedling length. This highlights the need for a genetic solution, encouraging breeders to prioritise the selection of oat genotypes with longer lengths adapted for hotter temperatures.

The field results showed no relationship between emergence from depth and oat genotypes’ combined lengths, as soil temperatures were not extreme and seeding depth did not exceed 110 mm, which is shorter than the total combined length of all oats genotypes tested. It is likely that in higher-strength soils, combined with warmer soil temperatures, having the genetic potential for a longer combined length would be beneficial to increase emergence ability. The possibility for obtaining such genetic traits from international oat genotypes is promising. International oat genotypes had combined mesocotyl and coleoptile lengths that were 20 mm greater than Australian genotypes and 15 mm greater than unreleased genotypes. This indicates that international genotypes might include useful genetics for increased combined length. Furthermore, breeders should target increased mesocotyl, as it exhibits higher heritability and greater genetic variation compared to coleoptile length and is also less affected by higher temperatures. Specifically, mesocotyl length was three times more genetically variable than coleoptile length, emphasising its potential as a key trait in breeding programs aimed at enhancing seedling emergence and early vigour. Fortunately, there was no correlation between mesocotyl and coleoptile lengths which indicates they are under separate genetic control. This variation presents a valuable opportunity for further research to identify and associate these traits with genetic markers or quantitative trait loci (QTLs) for deployment of simultaneous, multi-trait genomic selection (e.g. Meuwissen et al. 2001; Azizinia et al. 2023). By deploying appropriate genomic tools, breeders could develop more resilient oat genotypes with optimised mesocotyl and coleoptile lengths. Potential future studies could focus on mapping these traits to specific genetic regions, exploring the underlying genetic mechanisms, and evaluating the performance of selected genotypes in different environmental conditions.

Sowing early could expose winter cereals to frost during flowering, unless genotype phenology is matched so flowering time balances the risks of frost and heat stress and maximises the period for growth (Flohr et al. 2017; Cann et al. 2023). The increased risk of frost when flowering early is a relative advantage for oats. Oats demonstrate greater frost tolerance compared to wheat or barley. Oats tend to be more tolerant to frost than wheat and barley (DPIRD 2023b) and therefore, oats are a lower-risk option for farmers managing early sowing or frost-prone paddocks. This frost tolerance enhances their suitability for early sowing systems. Furthermore, the adoption of dual-purpose oats (genotype that can be both a grain and/or hay) offers further benefits, including increased profits, enhanced management flexibility, and improved risk mitigation. The early-sown dual-purpose oats give farmers flexible exit options, including grazing, haymaking, silage production, or grain harvest. This flexibility minimises risk exposure and maximises profit across diverse economic scenarios (Kirkegaard et al. 2019).

Limitations and future directions

The field experiment included 19 oat genotypes, four wheat genotypes, and two barley genotypes, which could introduce potential biases when comparing emergence. The small number of wheat and barley genotypes do not represent the full range of coleoptile lengths in these species. Nonetheless, the coleoptile lengths we observed were representative of DPIRD’s (2023a) assessments of modern wheat and barley genotypes. Wheat emergence presented correlated with the findings of Rebetzke et al. (2007). Although there is currently a lack of field data on emergence and biomass for deep-sown barley genotypes. Future research should aim to collect data on oats’ suitability to be deep-sown across a broader range of soil types in locations spanning different rainfall and climatic zones. Research investigating the impact of deep-sown oats on biomass and yield should also be done. Here, only data on emergence have been presented. Recent research by Zhao et al. (2023) and Niu et al. (2020) indicates that mesocotyl growth is more sensitive to chemical and environmental changes compared to the coleoptile during early growth stages. This heightened sensitivity suggests that the mesocotyl is more responsive to pre-emergent environmental conditions. Further research is required to identify oat genotypes better adapted to variable pre-emergence conditions.