Seed priming enhances early growth and improves area of soil exploration by roots

8 Introduction 9 Seed priming has been conducted for centuries with growth advantages reported for a variety 10 of different crops. Previous work has suggested priming does not offer a yield advantage 11 despite an increased early growth if grown under ideal conditions. However, how these 12 advantages unfold in regards to early root development is largely unknown. 13


Introduction 36
The process of seed germination can be divided into three steps; (1) imbibition, (2) activation 37 or lag phase and (3) root protrusion (Rajjou et al., 2012). In centuries of agricultural practise, 38 amendments were developed to improve the seed performance under varying conditions. 39 Physical seed enhancement technologies include magnetic (magnetic fluids used for removal 40 of contaminants), radiation (UV, microwave, ion radiation, X-ray and gamma-ray radiation 41 improve seed vigour but it is unclear how) and plasma (non-thermal plasma reduces pathogen 42 and chemical contamination in seeds) applications. Besides these physiological and physical 43 seed enhancement technologies, the chemistry of the seed can also be modified to increase 44 Historically, a variety of soaking methods have been reported that are affecting the 47 on the growth rate of seedlings in a growth comparison study between physical seed 95 enhancement technologies using X-ray CT (resolution 20 µm). 96 In general, little is known about the physical differences on early growth of primed seeds vs. 97 non-primed seeds. The aim of this study was to determine the influence of the priming process 98 on sugar beet seeds in terms of their in situ development and their early growth stage root 99 architecture. X-ray CT was used to non-destructively quantify the growth pattern of both 100 primed and non-primed seeds. 101

Treatment preparation 103
A loamy sand soil of the Newport series (83.2% sand, 4.7% silt, 12.1% clay and 2.93% organic 104 matter) was collected from the University of Nottingham farm at Bunny, Nottinghamshire, UK 105 (52.8586°, -1.1280°). Prior to packing, the soil was air-dried and sieved to < 1 mm. Sugar beet 106 (Beta vulgaris L.) seed material was supplied by Syngenta Seeds AB. Naked, untreated seeds 107 (NUS) were used alongside woodmeal and clay pelleted seeds (PUS) as well as seeds coated 108 with insecticide and fungicide additionally to the pelleting (PUSCO). Each treatment was 109 available as either primed (NS+, PS+, PS+CO) and non-primed treatment (NUS, PUS, PUSCO). 110 The naked coated treatment was omitted from this study as this treatment is not sold to the 111 end user and therefore of no collective interest. The seed pelleting and coating label, the 112 priming procedure, as well as the precise composition are treated confidentially. Four 113 replicates for each treatment were used in the study. 114 To compare differences in embryo and perisperm size between primed and non-primed seeds 115 an initial high resolution study was conducted on dry seed material outside of soil. Only naked 116 untreated seeds were used for this comparison as the priming treatment is conducted prior 117 to the application of physical enhancement technologies. Individual seeds were scanned using 118 a Phoenix Nanotom X-ray CT scanner (GE Measurement & Control Solutions, Wunstorf, 119 Germany) with an X-ray tube potential energy of 75 kV and a current of 120 µA. The detector 120 collected 1800 projection images (image average and skip were set to 3 and 1, respectively) 121 with a timing of 500 ms for each image. The scan spatial resolution and time were 2.5 µm and 122 The column packing was conducted as described in Blunk et al. (2017b). The soil columns were 126 scanned using a Phoenix v|tome|x m 240 kV X-ray CT scanner (GE Measurement & Control  127 Solutions, Wunstorf, Germany). The scans were conducted using an X-ray tube potential 128 energy of 130 kV and a current of 100 µA. The detector collected 2878 projection images with 129 timing of 250 ms per image (FAST SCAN mode; the sample continuously rotates during image 130 acquisition with no averaging or skip) at a resolution of 20 µm. To image the full length of the 131 column at maximum resolution, the 'multiscan' module in the acquisition software was used 132 to collect two scans per column resulting in a total scan time of 24 minutes (12 mins per 133 section). Reconstruction was conducted using the phoenix datos|x rec reconstruction 134 software with a beam hardening correction setting of 8 and an automatic calculation of the 135 region of interest and scan optimisation. All soil columns were scanned in the same order at 136 each time point to reduce temporal effects. 137

Soil core transplantation 138
Due to the design of the experiment, each soil core was transplanted to a larger column to 139 enable the highest possible resolution for all scanning days (day 2, day 4 and day 14 after 140 imbibition) as well as to allow enough room for the seedling to grow after day 4. The small 141 polypropylene column was pre-cut lengthways (secured with adhesive tape) and included 142 detachable mesh to enable a non-destructive extraction of the soil core following the first 143 stage of growth. After X-ray CT scanning following four days of growth, the soil core was 144 extracted from the column by detaching the mesh and opening the column along the 145 longitudinal axis (Fig. A.1). The soil core was then placed on top of a layer of 435 g dry soil with 146 a height of approximately half of the height of the large polypropylene column (170 mm height 147 and 76 mm inner diameter). The column was then filled with 405 g dry soil to generate a total 148 bulk density of 1.2 g cm -3 . The soil column was then saturated and drained afterwards to a 149 gravimetric moisture content of 20% w/w. Growth and moisture conditions were maintained 150 as previously described. 151 The larger soil columns were scanned using a Phoenix v|tome|x m 240 kV X-ray CT scanner 152 (GE Measurement & Control Solutions, Wunstorf, Germany) using an X-ray tube potential of 153 180 kV and a current of 180 µA. The detector collected 2399 projection images with timing of 154 column at maximum resolution, the 'multiscan' module in the acquisition software was used 157 to collect two scans per column resulting in a total scan time of 20 minutes (10 mins per 158 section). Reconstruction was performed as described earlier. The main effect of treatment was partitioned into the following contrasts: Contrast 1 allows 187 us to test the hypothesis that the pelleting procedure does improve radicle growth opposed 188 to naked seeds by testing NUS / NS+ against combined PUS / PS+ & PUSCO / PS+CO 189 measurements. If this contrast were to be significant then the implication is an improved 190 growth behaviour using pelleting technology despite application of coatings containing active 191 ingredients. Contrast 2 tests the hypothesis that the addition of a pesticide coating does 192 impact the growth behaviour by testing the comparison of PUS / PS+ against PUSCO / PS+CO. 193 If this contrast were to be significant, this would imply that the active ingredients in the 194 coating surrounding the pellet do have potential influence on the embryo development. If the 195 interaction of treatment and priming were to be significant, it would indicate that the chosen 196 treatment has an influence on the effect of priming. 197 The main effect of time was partitioned into the following contrasts: Contrast 1 test the 198 hypothesis that the radicle growth in the treatments is of linear nature. If this contrast were 199 to be significant then the implication is a constant growth rate throughout the measured time 200 interval. Contrast 2 allows us to test the hypothesis that the effect of time is on non-linear 201 nature which would imply a non-constant growth if this contrast were to be significant 202 (Wishart and Metakides, 1953). For this instance, orthogonal contrasts for unequal intervals 203 have been used (Snedecor, 1958). If any interaction with the factor time were to be significant,

Results 210
Prior to the growth comparison of primed and non-primed seeds, morphological differences 211 between the treatments were quantified by scanning seeds ex situ. The embryo dimensions 212 were assessed as one object as differences in greyscale levels between the organic parts were 213 low and did not allow a distinct separation between cotyledons, hypocotyl and root (Figure 1). 214 water filled pores, organic matter and mineral grains. C) 3D rendered X-ray CT image of a 218 naked seed in soil. 219 No significant differences were found between the primed and non-primed seeds for embryo 220 volume and surface area as well as perisperm and surface area (Table 1). However, we noted 221 a trend of increased embryo volume and surface area as well as perisperm surface area in the 222 primed treatment compared to the non-primed treatment (Table 1). 223 226 The growth behaviour of the seedlings was quantified over time and the transplantation 227 method appeared successful for all replicates without disturbing the soil matrix. Initial 228 statistical analysis was performed as a factorial linear mixed effect model and the individual 229 effects analysed using orthogonal contrasts as displayed in Table 2. 230   Figure 4B) and number of laterals ( Figure 4A). However, no significant effect of priming 265 was found with regards to the average lateral root length (p = 0.10) ( Table A.2). The difference 266 in root architecture was furthermore quantified using the convex hull (smallest convex object 267 set containing all roots). A significantly larger convex hull was observed for the treated seeds   interactions between the contact area and growth characteristics (e.g. tap root length and 284 lateral growth) as well as treatment information (e.g. priming or pelleting) due to a high 285 variability within the dataset. A regression between the seed-soil contact on day 2 and the 286 root growth rates was fitted separately (Fig. A.2). In general, seed-soil contact did not 287 correlate with growth rate at any of the three time points measured (p = 0.54). The R 2 values 288 for all treatments showed a low conformity of the fitted regression line to the data points 289 except for PS+ on day 4 with an R 2 of 0.80 showing an increased growth rate with rising contact 290 area. A negative trend was observed for NS+ on day 2 with an R 2 of 0.63 exhibiting a decrease 291 growth rate with increasing seed-soil contact. 292

Discussion 293
A common assumption of the seed priming process is that biological processes are initiated 294 inducing all metabolic activities necessary for germination, however almost no morphological 295 differences occur as they are irreversible (Hill, 1999). Although no significant differences were 296 observed in the volume and surface area of seed embryo and perisperm in this study, a 297 positive trend was detected that could suggest swelling of these structures during the 298 germination process in primed seeds which was also described earlier for Parsley 299 The primed seeds had a significantly faster growth rate over the first four days compared to 305 non-primed seeds which agrees with previous findings stating a uniform and accelerated 306 germination using varying priming techniques (Paparella et al., 2015). Furthermore, as 307 significant combined effects were found for priming, treatment and day using the linear mixed 308 effect model it highlights the growth advantages of seed priming regardless of the applied 309 physical enhancement. In general, the utilisation of seed storage reserves diminishes upon 310 seedling growth by a shift from a hetero-to an autotrophic metabolism (Bewley and Black, 311 1994). The direct impact of seed pelleting applications on seedling growth is therefore 312 disrupted upon disconnection of the seedling from the seed transitioning to a soil nutrient-313 based growth which was observed for most of the seedlings between day 2 and day 4 after influence on the seedling. Primed seeds have been reported to have a similar ultimate yield 316 under ideal conditions compared to a non-primed treatment supporting our observation of 317 similar tap root lengths after 14 days of growth which might also be an artefact of restricted 318 growth due to the vessel size (Danneberger et al., 1992). Also, the number of basal roots was 319 reported as being similar for pepper after 14 days of growth agreeing with our findings of no 320 significant differences in number of lateral roots upon priming (Stoffella et al., 1992). 321 Furthermore Leskovar and Cantliffe (1993)