Scholarly article on topic 'Co-ordination between primordium formation and leaf appearance in soybean ( Glycine max ) as influenced by temperature'

Co-ordination between primordium formation and leaf appearance in soybean ( Glycine max ) as influenced by temperature Academic research paper on "Biological sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Field Crops Research
OECD Field of science
Keywords
{Co-ordination / Soybean / " Glycine max " / Phyllochron / Plastochron}

Abstract of research paper on Biological sciences, author of scientific article — Fatima M. Tenorio, James E. Specht, Timothy J. Arkebauer, Kent M. Eskridge, George L. Graef, et al.

Abstract Soybean (Glycine max) production is expanding into cooler and warmer environments. Temperature influence on nodal leaf appearance rate and, especially, nodal primordium formation rate – two critical parameters that determine potential leaf area, light absorption, and crop growth and yield – is poorly understood. This study was designed to determine the influence of temperature on primordium formation and leaf appearance and to investigate how the co-ordination between these two processes is affected by contrasting temperature regimes. The experiments were conducted in field and greenhouse settings using indeterminate cultivars differing in maturity group. Soil and air temperature was measured at 30-min intervals in all experiments, starting at sowing and ending at physiological maturity. Plants were dissected every 4–7d to determine the number of primordia at the stem apical meristem and to estimate primordium formation rate. The number of nodal leaves at main stem was assessed periodically during the entire growing season to estimate rate of leaf appearance. Primordium formation ended near the beginning of pod setting, while leaf appearance ceased at the beginning of seed filling. The end-season final number of primordia was greater than the final leaf number, revealing a surplus of primordia that did not advance beyond primordial stage. Across experiments, mean temperature during the phases of primordium formation and leaf appearance ranged from 15 to 26°C. Both primordium formation and leaf appearance were temperature-dependent, but primordium formation was faster than leaf appearance. The plastochron was 36°Cd, whereas the phyllochron was biphasic, decreasing from 83°Cd to 58°Cd during ontogeny. A strong relationship was found between the number of primordia and the number of leaves, which was stable across experiments, temperature treatments, and cultivars. This study has established that both primordium formation and leaf appearance in soybean are influenced by temperature, with a co-ordination between these two processes. The co-ordination model presented here can be used for robust prediction of seasonal nodal leaf dynamics in soybean.

Academic research paper on topic "Co-ordination between primordium formation and leaf appearance in soybean ( Glycine max ) as influenced by temperature"

Contents lists available at ScienceDirect

Field Crops Research

journal homepage: www.elsevier.com/locate/fcr

Original Article

Co-ordination between primordium formation and leaf appearance in soybean (Glycine max) as influenced by temperature

CrossMarlc

Fatima M. Tenorioa, James E. Spechta, Timothy J. Arkebauera, Kent M. Eskridgeb, George L. Graefa, Patricio Grassinia'*

a Department of Agronomy and Horticulture, University of Nebraska-Lincoln, Lincoln, NE 68583-0915, USA b Department of Statistics, University of Nebraska-Lincoln, Lincoln, NE 68583-0963, USA

ARTICLE INFO

Keywords:

Co-ordination

Soybean

Glycine max

Phyllochron

Plastochron

ABSTRACT

Soybean (Glycine max) production is expanding into cooler and warmer environments. Temperature influence on nodal leaf appearance rate and, especially, nodal primordium formation rate - two critical parameters that determine potential leaf area, light absorption, and crop growth and yield - is poorly understood. This study was designed to determine the influence of temperature on primordium formation and leaf appearance and to investigate how the co-ordination between these two processes is affected by contrasting temperature regimes. The experiments were conducted in field and greenhouse settings using indeterminate cultivars differing in maturity group. Soil and air temperature was measured at 30-min intervals in all experiments, starting at sowing and ending at physiological maturity. Plants were dissected every 4-7 d to determine the number of primordia at the stem apical meristem and to estimate primordium formation rate. The number of nodal leaves at main stem was assessed periodically during the entire growing season to estimate rate of leaf appearance. Primordium formation ended near the beginning of pod setting, while leaf appearance ceased at the beginning of seed filling. The end-season final number of primordia was greater than the final leaf number, revealing a surplus of pri-mordia that did not advance beyond primordial stage. Across experiments, mean temperature during the phases of primordium formation and leaf appearance ranged from 15 to 26 °C. Both primordium formation and leaf appearance were temperature-dependent, but primordium formation was faster than leaf appearance. The plastochron was 36 °Cd, whereas the phyllochron was biphasic, decreasing from 83 °Cd to 58 °Cd during ontogeny. A strong relationship was found between the number of primordia and the number of leaves, which was stable across experiments, temperature treatments, and cultivars. This study has established that both pri-mordium formation and leaf appearance in soybean are influenced by temperature, with a co-ordination between these two processes. The co-ordination model presented here can be used for robust prediction of seasonal nodal leaf dynamics in soybean.

1. Introduction

Leaf number at a given point of time depends upon the rate of primordium formation (microscopically assessed) at the stem apical meristem (SAM) and the rate of (macroscopic) leaf appearance. Temperature is the main environmental factor determining rates of primodium formation and leaf appearance in annual crop species (Granier and Tardieu, 1998; Kiniry et al., 1991; Sadras and Villalobos, 1993). Typically, these rates linearly increase over a range of

temperatures, delimited by a base temperature (Tb), below which rate becomes zero, and by an optimal temperature (Topt), above which rates decrease (Kiniry et al., 1991; Slafer and Rawson, 1994). The inverse of the slope of the linear relationship between primordium formation and leaf appearance rates plotted versus temperature represents the plastochron (defined as thermal time [°Cd] between two successively formed primordia) and the phyllochron (the thermal time between two successively appeared leaves), respectively. The plastochron and phyl-lochron are relatively constant across management practices and

Abbreviations: GH, greenhouse; HT, high temperature; IT, increasing temperature; LR, nodal leaf rate; LT, low temperature; MG, maturity group; MT, moderate temperature; NL, nodal leaf; NP, nodal primordium; NLMAX, end-season maximum nodal leaf number; NPMAX, end-season maximum primordium number; PR, nodal primordium rate; R1, beginning of flowering; R2, full flowering; R3, beginning of pod setting; R5, beginning of seed filling; R7, physiological maturity; SAM, stem apical meristem; SD, sowing date; VE, emergence; Tb, base temperature

* Corresponding author. E-mail address: pgrassini2@unl.edu (P. Grassini).

http://dx.doi.org/10.10167j.fcr.2017.03.015

Received 23 January 2017; Received in revised form 30 March 2017; Accepted 31 March 2017

Available online 16 June 2017

0378-4290/ © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

environmental conditions (Kiniry et al., 1991; Ritchie and NeSmith, 1991; Slafer, 1995), although genotypic variation has been reported (Padilla and Otegui, 2005; Slafer and Rawson, 1997).

With soybean production (Glycine max) currently expanding into cooler latitudes and into warmer environments (Sinclair et al., 2013; Specht et al., 2014), it would be useful to bolster current knowledge about the influence of temperature on soybean primordium formation and leaf appearance. The appearance of each successive leaf influences the seasonal leaf area index pattern (Setiyono et al., 2011), which, in turn, controls light absorption, dry matter production, and seed yield determination (Board, 2004; Egli and Zhen-wen, 1991; Sinclair, 1984b). In the present study, we used the terms nodal primordium rate (PR, primordia d-1) and nodal leaf rate (LR, leaf d-1). A nodal primordium was declared to be "new" when its size was microscopically observed to have attained a length of 80 |jm (Sun, 1957). A nodal leaf was declared to have "appeared" when the size of that leaf was mac-roscopically observed to have attained a length of 21 mm (Sinclair, 1984b). The first nodal leaf corresponds to the node bearing the cotyledons. In determinate soybean cultivars plants, the successive appearance of nodal leaves ceases when the terminal bud becomes a raceme (Caffaro and Nakayama, 1988; Carlson, 1973). This transition is less abrupt in indeterminate cultivars, with leaf appearance eventually ceasing at the beginning of seed filling (R5 stage; Fehr and Caviness, 1977) and without presence of a terminal raceme (Sinclair, 1984a; Saitoh et al., 2004; Bastidas et al., 2008).

Temperature influence on soybean LR was first documented in the early 1970s by Hesketh et al. (1973). These researchers grew soybean indeterminate cv Wayne (maturity group [MG] 3.0) in a series of greenhouse and growth chamber experiments where plants were exposed to specific temperatures that ranged from 13 to 30 °C. That study documented a linear relationship between LR and temperature, and the phyllochron was estimated to be 56 °Cd for each successive nodal leaf, with a Tb of 6 °C (Hesketh et al., 1973). A few years later, Thomas and Raper (1976) reported an identical phyllochron for determinate cv Ransom (MG 7.0), but with a slightly higher Tb (9 °C), based on experiments conducted in growth chambers, set to specific temperatures that ranged from 16 to 28 °C. However, it was not clear if the temperature used in the two studies were measured at SAM height, or just the targeted (i.e., temperature set in greenhouse or growth chamber). Phyllochron estimates reported in these studies have not yet been validated in field-grown soybean experiments. Recently, Bastidas et al. (2008) concluded that LR was insensitive to temperature, reporting a near-constant LR of 0.27 leaves d-1 (i.e., 3.7 days leaf-1) across factorial set of 14 cultivar x four sowing date treatments in field experiments conducted over two years. However, the limited spring temperature range (22-26 °C) in that experiment was likely insufficient for assessing the influence of temperature on LR. Moreover, the LR tem-perature-insensitivity observed in that study and the LR temperature-sensitivity observed in the two prior studies are conflicting.

In these three prior studies, soybean nodal primordium formation was not evaluated nor was the influence of temperature on PR examined. Miksche (1961) reported a PR of 0.29 primordia d-1 (i.e., 3.5 days between successive primordia) from the first to the second trifo-liolate primordia, but thereafter observed a constant PR of 0.5 pri-mordia d-1 (i.e., a 2-day interval between successive primordia), but unfortunately, did not report the corresponding air temperatures. Johnson et al. (1960) observed that just 35 days after sowing, the SAM had already produced all of the 19 primordia that eventually appeared as leaves. Based on this 1960 report, Lersten and Carlson (2004) inferred a PR of 0.5 primordia d-1 (i.e., 2 days). Thomas and Kanchanapoom (1991) and Chiera et al. (2002) reported respective PR values of 0.5 primordia d-1 for a 26/22 °C day/night temperature regime, and 0.52 primordia d-1 for a 24/22 °C day/night temperature regime. A strong synchronization between PR and LR (hereafter called "co-ordination model") has been reported for many crop species in the literature. Kirby (1990) and Hay and Kirby (1991) observed the

existence of a strong relationship between primordium formation and leaf appearance in wheat, which was not affected by genotype or photothermal conditions. The interdependence between primordium formation and leaf appearance under contrasting temperature regimes has not been investigated in soybean.

Reliably accurate prediction of nodal leaf dynamics across a range of weather, management, and cultivars is crucial for accurate simulation of soybean phenology, leaf area, and seed yield. However, as noted, little is known about the influence of temperature on PR and LR and the nature of the co-ordination between these two processes in indeterminate soybean, in which the SAM continues to differentiate new primordia as floral differentiation progresses from the lower to the upper nodes. To bridge this knowledge gap, the aim of this study was to investigate the response of PR and LR to temperature in indeterminate soybean cultivars in a combination of experiments conducted in controlled environments and field conditions.

2. Materials and methods

2.1. Greenhouse experiments

The experimental design used in the greenhouse (GH) experiments was replicated twice in time with a split-plot randomized complete block design in which the replicates were the spring and fall GH seasons of 2015 (Table 1; Supplementary material, Table S2). Plants were grown in 21-L plastic pots (0.3 m height and diameter) filled with sandy loam soil. Pots were placed in four GH bays ('main plots'), simulating a traditional plant density of 36 plants m- 2, with four plants per pot set to be 0.15 m apart. A total of 64 pots (i.e., 256 plants) were allocated to each of the four GH bays and each GH pot set was surrounded by border pots of plants. To account for any East-West temperature gradient in each GH bay, the 64-pot sets were arrayed into four E-W 16-pot subsets ('blocks') for sub-sampling measurement purposes. Twelve pre-germi-nated seeds (radicle = 5 mm) were initially sown per pot on 21 February 2015 (spring GH replicate) and 16 September 2015 (fall GH replicate) and were thinned to four plants per pot at emergence (VE stage; Fehr and Caviness, 1977), resulting in 32 plants per block for each cultivar MG in each GH temperature treatment. Plants were kept well watered using a drip irrigation system, with amounts of water that were periodically adjusted to match seasonal variation in crop water demand. Plants were fertilized with a total dressing of 1.5 g N, P and K per pot, applied shortly after VE stage. Plants received periodic prophylactic applications of fungicide and insecticide.

The four GH bays were randomly assigned in each GH replicate to four different day/night temperature regimes applied during the growing season. In three of the four treatments, the day/night temperatures did not change during the season and were characterized as high (HT; 28/20 °C), moderate (MT; 23/15 °C), or low (LT; 18/10 °C), differing from each other 5 °C in both day and night. The remaining treatment (increasing temperature [IT]) was designed to mimic the typical increase in air temperature that occurs from spring to summer between VE to R5 stages in field-grown soybean in major producing areas, such as the US Corn Belt and South American Pampas. In that treatment, the day/night temperature was increased by 2 °C each week from 18/10 °C at V3 stage to 28/20 °C at R5. Artificial extension of the natural solar photoperiod to 15 h in all GH treatments was accomplished by using incandescent lamps from sowing to maturity. The fall GH replicate differed somewhat from spring GH replicate in that: (i) solar radiation decreased over time in the fall whereas is increased over time in the spring; (ii) non pre-germinated seeds were sown due to the lack of difference in dates of VE between pre-germinated and non-pre-germinated seeds observed in spring GH replicate tests; and (iii) two treatments (IT and LT) unfortunately had to be terminated around 85 days after sowing (just before R1 stage, which corresponds to beginning of flowering; Fehr and Caviness, 1977) due to a severe uncontrollable powdery mildew infestation caused by fungus Microsphaera diffusa.

Table 1

Main features of the four experiments conducted during 2015 to assess the influence of temperature on formation of nodal primordia and appearance of nodal leaves in soybean.

Experiment Location (latitude and longitude)

Condition

Growing season

Soil type

Cultivar name (and MG)

Spring GH Lincoln, NE (40.80 N, 96.68 W)

Fall GH Lincoln, NE (40.80 N, 96.68 W)

SD expt. Lincoln, NE (40.80 N, 96.68 W)

Farmer field Atkinson, NE (42.47 N, 98.75 W)

Saronville, NE (40.57 N, 98.13 W) Smithfield, NE (40.58 N, 99.67 W) Mead, NE (41.15 N, 96.48 W)

Greenhouse

Greenhouse

Field experiment

Spring (Feb -May)

Fall (Sept-Dec)

Field experiment Summer (April-Oct)

Summer (April -Oct)

Sandy loam

Sandy loam

Silt loam

Sandy loam Silty clay loam Silt loam Silty clay loam

IA 2102 (MG 2.1) IA 3024 (MG 3.0)

IA 2102 (MG 2.1) IA 3024 (MG 3.0)

IA 2102 (MG 2.1) IA 3024 (MG 3.0)

A2733 (MG 2.7) A2431 (MG 2.4) P24T19 (MG 2.4) P31T11 (MG 3.1)

Abbreviations: HT: high temperature; MT: moderate temperature; IT: increasing temperature; LT: low temperature; SD: sowing date; F: farmer; GH: greenhouse; MG: maturity group.

2.2. Field experiments

The purpose of the sowing date field experiment (SD expt.) was to expose field-grown soybean to different mean temperature during the phases of primordium formation and leaf appearance (Table 1; Supplementary material, Table S2). The SD expt. was replicated (i.e., four blocks) and followed a split-plot randomized block design. The main plots were six sowing dates that successively differed by ca. 10-day intervals, with the first and last sowing on 23 April 2015 and 19 June 2015, respectively. The subplots were the two cultivars (of MG 2.1 and 3.0). The six-row subplot row length was 4.6 m, with an inter-row spacing of 0.76 m, resulting on a subplot area of ca. 21 m2. Seeds were sown at 2.5 cm spacing (39 plants m-2) at a 6-cm sowing depth, then thinned to 30 plants m-2 at VE. The last row for each side of each subplot served as border rows. The measured seed yields, expressed at 13% seed moisture content, ranged from 3.5 (late sowing) to 4.6Mgha-1 (early sowing). A pre-emergent herbicide (Dual II Magnum) was applied before sowing and any weeds that emerged during the growing season were manually removed. Irrigation was applied with a drip irrigation system and scheduled as needed to replenish soil water content after adjusting for local rainfall events.

Additional field experiments were conducted in four farmer irrigated fields located in Nebraska. These four fields provided a wider range of temperature and weather, soil, and management practices in soybean fields across N-S and E-W Nebraska soybean production zones for evaluating the effect of temperature on PR and LR (Table 1). Each field provided a sampling site, though in each field four separate blocks were evaluated. Each 22-row block had 0.76 m inter-row spacing, and was 16.8 m wide and 19.1 m in length. The viable seeding rate was 27-34 plants m - 2. Fields were sown and managed by farmers. Sowing date varied across the four fields as a result of differences in early-season weather. Cultivars differed but were the recommended ones used in the region where fields were located. Crops were watered frequently by farmers to maintain available soil water content above 65% of soil plant available water holding capacity in the 0-90 cm soil depth throughout the season. Fields received applications of 19-30 kg ha-1 P along with 4-67 kg ha -1 N (either as a direct fertilizer or a credit from the nitrate in the irrigation). Farmers applied pre- and post-emergence herbicide to control weeds and also applied foliar fungicide and insecticide at the beginning of pod setting (R3 stage; Fehr and Caviness, 1977). Measured end-season seed yield ranged from 5.1 to 5.9 Mg ha-1. There was no detectable influence of pesticide application on leaf appearance and primordium formation dynamics.

2.3. Identification of new nodal primordia and new nodal leaves

Nearly all soybean researchers use a vegetative and reproductive staging system published about 40 years ago to describe main stem node development (Fehr and Caviness, 1977) (Supplementary material, Table S1). In this staging system (which we will hereafter refer to as FC for brevity), the cotyledon node is considered to be node zero, and the uppermost node counted at any given time is the node just below a node whose leaf has just unrolled. We also used the FC system in this study, but used a modification of that system to track the leaf number (NL), wherein considered the cotyledon node to be NL1 and also counted as the uppermost node the node that had a just unrolled leaf (not the node below, which is done in the FC system). Sinclair (1984b) noted that unrolled leaflets had a length of 21 mm, which we corroborated in our study. Our system was thus not fundamentally different from the FC system; in effect, our NL is simply always two more than the FC vegetative node number (e.g., FC V2 is equivalent to our NL4). No attempt was done to track leaf appearance and primordium formation in branches. We note, however, that soybean plants grown in our experiments had very few branches as a result of the high plant population density, which is the common practice in US soybean farmer fields.

Plants were staged every 2 d, except that a 7 d interval was used for the four distant farmer fields, from sowing until cessation of leaf appearance. In the spring and fall GH replicates, eight plants located in the middle of each block were staged for each MG cultivar. In SD expt. and four farmer fields, ten contiguous plants located in one of the center rows in each block were used for staging. The seasonal timing of each successive primordium was determined using destructive microscopic sampling of the SAM. The SAM samples were collected every 4 d (GH replicates and SD expt.) and every 7 d in farmer fields from sowing until cessation of primordium formation at the main SAM. The SAM samples were collected from plants located in the center rows of the blocks. After a SAM sample was collected, it was immediately placed in 8 to 12 mL glass vials, containing a standard solution of formaldehyde (100 mLL-1), ethyl alcohol (500 mLL-1), glacial acetic acid (50 mL L-1), and distilled water (350 mL L-1). The preserved SAM sample was later dissected in the laboratory using a binocular microscope (Nikon SMZ-10) at 10-60X magnification. The number of pri-mordia (NP) was counted starting from the cotyledonary node up to the last formed primordium on the flank of the SAM. Following Sun (1957), a primordium was considered formed when it had attained a length of ca. 80 |jm. Coefficient of variation for NP was very small (< 5%) across sampling dates in all experiments and treatments.

Treatment

Soil and air temperature was measured at 30-min intervals in all experiments, starting at sowing and ending at physiological maturity (R7 stage; Fehr and Caviness, 1977). Soil temperature sensors were placed at seed depth, whereas air temperature sensors were continuously adjusted to be at SAM height. Average daily temperature was calculated by averaging the (48) temperature records in a 24-h day. In all calculations, daily soil temperature was only used for the time period when the SAM was below soil surface (i.e., from sowing to NL1); daily air temperature was used thereafter.

2.4. Analysis of dynamics of nodal primordia and leaves

The apparent tri-phasic pattern of NL was fit to a tri-segment linear regression model:

Nl = a + Lri x t for t < Xi (1)

Nl = a + Lri x t + Lr2 x (t - Xi), for Xi < t < X2 (2)

Nl = Nlmax for t > X2 (3)

where a is the intercept of the first phase, LR1 and LR2 are the leaf appearance rates (leaves d-1) corresponding to the first (Eq. (1)) and second (Eq. (2)) phases, t is the number of days after sowing, X1 and X2 are the breakpoints separating the initial and subsequent linear phases (d), and NLMAX is the end-season maximum NL. The model was fitted to the observed data from NL1 until R7 stage. Model fitting was implemented with GraphPad Prism (GraphPad Software® v. 6.07). Fitting the tri-segment linear model was not possible in fall GH, due to an unstable GH temperature setting during the early vegetative stages and the forced early termination of the experiment due to a severe powdery mildew infestation. Thus, LR1 for the treatments in the fall GH was computed as the quotient between the difference in NL accrued between NL1 and NL4 (i.e., three nodal leaves) and the number of days between the appearance of NL1 and NL4. A similar approach was used to compute LR2 between the appearance of NL4 and NLMAX.

The seasonal pattern of NP seemed to be linear in all experiments; hence, a simple linear function was fitted to that data:

Np = Nps + Pr x t (4)

where NPS equals two to represent the fact that the seed already has the cotyledon and the unifoliolate nodal primordium, PR is the pri-mordium formation rate (primordia d-1), and t is the number of days after sowing. A two-segment linear regression model (i.e., two phases with different slopes) provided a better fit for the IT treatments in spring and fall GH experiments. PR for IT treatment was calculated as the average of the two slopes weighted by the phase length associated with each slope.

2.5. Temperature-based phyllochron and plastochron estimation

The parameters LR1 and LR2 were derived from the estimated regression coefficients in Eqs. (1) and (2), while PR was derived from the estimated regression coefficient in Eq. (4). For each treatment, a mean temperature was calculated for the time period between NL1 and X1 (LR1), between X1 and X2 (LR2), and between sowing to end of pri-mordium formation (PR). No statistically significant difference in the estimated slope and Tb for the relationships between PR and LR versus temperature was observed when the GH and field experimental data were analyzed separately (P >0.30). Likewise, no statistically significant difference was detected between the MG 2.1 and MG 3.0 cul-tivars for those same parameters (P > 0.44). Therefore, data across treatments and experiments were pooled and a generic linear regression model was fitted to model the observed response of PR and LR to mean temperature. The LR data for MG 2.1 from HT and MT treatments in fall GH were excluded from the LR2 regression analysis because of the short time period between X1 and X2, which made unfeasible a reliable estimation of LR2.

Plastochron and phyllochron (°Cd) were estimated by taking the inverse of the slope of the linear model fitted to the relationship between PR or Lr and mean temperature. The Tb parameter for PR and LR was estimated by extrapolating the fitted regression equations to their respective x-intercept values.

2.6. Co-ordination between nodal primordium formation and leaf appearance

A two-segment linear relationship between NP and NL during the course of the growing season was evident after visual inspection of the data. Statistical analysis indicated that the two-segment model was much more likely than a simple linear model for modeling the observed trend (F-test, P < 0.001). Hence, a two-segment linear model was fitted to regress the NL on the NP:

Nl = b + c x NP, for NP < d (5)

Nl = b + c x d + e x (NP - d), for NP > d (6)

where b is the intercept of the first phase, c and e are the slopes of the first (Eq. (5)) and second (Eq. (6)) phases, and d is the breakpoint between the two phases. Data on NL after reaching the maximum NP (NPMAX) were not used when the two-segment model was fitted. The two-segment model was fitted separately to the data from each experiment. The parameters of the model did not significantly differ among experiments (P > 0.15) or experimental conditions (P > 0.45), so the data were pooled across experiments. A single generic two-segment linear model was then fitted to the pooled data. Finally, a small dataset on NP and NL, reported as an average of six cultivars (i.e., determinate and indeterminate) by Johnson (1960), was added to the plot for comparison against the data collected in the present study.

Treatment effects on measured and calculated variables were evaluated by analysis of variance (ANOVA) performed separately for each experiment (Supplementary material, Table S6-S8). A Tukey test was used to determine significant differences (a = 0.05) between treatment means.

3. Results

3.1. Environmental conditions

Experiments conducted in the field in this study encompassed a range of weather, soils, and management practices, and of course both the field and GH experiments were designed to expose the plants to a range of temperatures, which influenced the duration of the time intervals amongst phenological stages (Fig. 1; Supplementary material, Tables S2 and S3). For example, the time period between sowing to R5 ranged from 52 d (fall GH, MG 2.1, HT) to 101 d (spring GH, MG 3.0, LT). Plants were exposed to a range of photo-thermal conditions during the periods of primordium formation and leaf appearance; mean temperature and incident radiation ranged from 15 to 26 °C and 5.9 to 22.1 MJm-2d-1, respectively (Fig. 1; Supplementary material, Table S3). The range of temperature explored in our study was representative of the range of mean temperature to which soybean crops are exposed in most soybean producing environments. Our temperature range was also consistent with that reported in previous studies aimed to determine plastochron and phyllochron in other crop species (e.g., Padilla and Otegui, 2003; Sadras and Villalobos, 1993). In contrast, the range in mean photoperiod (considering a solar elevation >—6.00° to account for the twilights) was narrow (15-16.3 h) across treatments and experiments (Fig. 1; Supplementary material, Table S3).

3.2. Seasonal dynamics of nodal primordia and leaves

A tri-segment linear model provided the best fit (R2 > 0.99; P < 0.001) for the observed seasonal patterns of leaf appearance

Fig. 1. Seasonal patterns in average daily air temperature (°C) and photoperiod (P; h) in the greenhouse (GH) experiments (spring and fall replicates), in the field sowing date experiment (SD expt.), and in four farmer fields (Fn). Dotted lines indicate mean temperature for each treatment in spring and fall GH (note that a moving weekly mean temperature starting at V3 stage was calculated for IT). Horizontal lines at the bottom of each panel indicate duration from sowing to the date of beginning of seed filling (R5 stage; Fehr and Caviness, 1977) for each treatment, with dashed lines indicating the longer duration of MG 3.0 in all experiments except for farmer fields. See Table 1 for treatment codes and description.

Fig. 2. Seasonal patterns in nodal leaf number for all treatments across the greenhouse (GH) experiments (spring and fall replicates), sowing date field experiment (SD expt.), and farmer fields (Fn). Symbols indicate the maturity group (MG) in GH and SD experiments. Arrows indicate date of beginning of seed filling (R5 stage; Fehr and Caviness, 1977) for MG 2.1 (f) and MG 3.0 (4) in GH and SD expt. Tri-segment linear regression models were fitted to all treatments from emergence until the cessation of leaf appearance, except for fall GH. Estimates of model parameters are shown in Supplementary material, Table S4. See Table 1 for treatment codes and description.

(Fig. 2; Supplementary material, Table S4). The duration of the first phase of leaf appearance, first breakpoint (Xi), ranged from 18 to 49 d after sowing (DAS) across treatments (Supplementary material, Table S4). The duration of the period between NL1 and X1 (range: 6-46 d) was not associated with air temperature (r2 = 0.01; P = 0.66). The estimated breakpoint X1 value varied among treatments, occurring as soon as NL2 but no later than NL6 (Fig. 2). For the same temperature, or same sowing date treatment in the spring and fall GH replicates, or SD expt., the MGs 2.1 and 3.0 cultivars exhibited almost identical LR1 and X1 values (Supplementary material, Table S4).

The second phase of leaf appearance commenced at breakpoint X1 and continued until the cessation of leaf appearance (Fig. 2). The date at which the last leaf appeared (X2) ranged from 45 to 101 DAS, depending on the treatment, but leaf appearance cessation DAS coincided approximately with the R5 stage (Fig. 2; Supplementary material, Tables S2 and S4). The abrupt decline in leaf appearance upon the approach to R5 observed in our study (Fig. 2) was previously noted by both Setiyono et al. (2007) and Sinclair (1984a). The NLMAX at R5 ranged from 6 (fall GH, HT&MT, MG 2.1) to 19 (SD expt., SD3, MG 3.0 & farmer field F4) across treatments (Supplementary material, Table S4). For a given temperature, or for given sowing date treatment in spring and fall GH replicate and SD expt., the MG 2.1 and 3.0 cultivars exhibited almost identical LR2 values (except for the spring GH, & HT). The MG 3.0 cultivar had a longer X1 to X2 duration because of its later R5 date, and thus it finished the season with a higher NLMAX (Fig. 2; Supplementary material, Table S4).

In contrast to the tri-phasic seasonal NL pattern, the seasonal NP pattern was quite linear, and thus a simple linear regression model (r2 > 0.95; P < 0.001) provided the best fit (Fig. 3; Supplementary material, Table S5). The two exceptions were the better fits of a two segment linear model for the MG 3.0 cultivar in IT treatment in spring GH (F-test, P < 0.001) and fall GH (F-test, P = 0.006). Three common

features were notable in the observed primordium formation patterns across all experiments. First, two primordia (i.e., cotyledonary and unifoliolate) were already formed in the quiescent seed and a trifolio-late primordium was already microscopically observable, though had not yet attained a length of 80 |jm (Figs. 3, 4A, B). Second, NPMAX was attained slightly after 'full flowering' stage (R2 stage; Fehr and Caviness, 1977), but clearly before R3 stage in all treatments (Figs. 3, 4H; Tables S2 and S5). Third, the NPMAX always exceeded the NLMAX in each treatment (Figs. 2 and 3; Supplementary material, Tables S4 and S5), with magnitude of this surplus ranging from 1 (farmer field F3) to 10 primordia (spring GH, MG 3.0, IT&MG 2.1, LT). Thus, several late-formed main stem primordia never became fully developed leaves at the stem apex before the crop matured.

3.3. Temperature influence on nodal primordium and nodal leaf rates

Temperature had a strong influence on soybean LR and PR, with faster rates with increasing mean temperature (Figs. 2 and 3; Supplementary material, Table S4-S8). However, PR was much lower in HT treatments in greenhouse experiments relative to field treatments that explored the same temperature range, probably due to an interaction of very low radiation and a higher than expected temperature in the GH (Supplementary material, Table S3). A lower PR in high temperature-low radiation greenhouse setting has also been reported for sunflower (Sadras and Villalobos, 1993) and wheat (Rawson, 1993; Rawson and Zajac, 1993). Given the unlikely scenario of a season-long period of very low radiation coupled with high temperature ever occurring in field-grown soybean during the entire crop season, the PR data from HT in greenhouse experiments were treated as outlier data and were excluded from the regression analysis.

The modeled relationship between PR and LR2 with temperature, using the pooled data, accounted for most of the observed variability

Fig. 3. Seasonal patterns in nodal primordium number for all treatments across the greenhouse (GH) experiments (spring and fall replicates), field sowing date experiment (SD expt.), and farmer fields (Fn). Insets show patterns of NP formation for IT treatments in spring and fall GH replicates. Symbols indicate maturity group (MG) in spring GH, fall GH, and SD expt. Arrows indicate date of beginning of pod setting (R3 stage; Fehr and Caviness, 1977) for MG 2.1 (f) and MG 3.0 (j) in spring GH, fall GH, and SD expt. Linear regression lines were fitted following Eq. (4). Estimates of model parameters are shown in Supplementary material, Table S5. See Table 1 for treatment codes and description.

Fig. 4. A germinated soybean seed showing: (A) the cotyledon (CT) and unifoliolate nodal primordium (UP) under 15x magnification, and (B) the stem apex meristem (SAM) at higher magnification (60x). Note that a trifoliolate nodal primordia (TP) was already microscopically observable, though had not yet attained a length of 80 A soybean terminal bud at emergence (with cotyledons removed) showing: (C) the UP under 10x magnification; (D) the first (1) and second (2) formed TP under 45x magnification; (E) the SAM, with the third (3) TP under 60x magnification, and (F) the three TP alternately formed at the SAM under 45x magnification. A soybean terminal bud: (G) before beginning of pod setting (R3 stage; Fehr and Caviness, 1977) showing the TP nodal primordia formation at the flank of SAM. Axillary buds have already formed a floral primordia (FP), and (H) after R3 showing differentiation of FP and bracts (BR) in the axillary bud adjacent to the SAM, coincident with the end of nodal primordia formation. S = stipule.

Fig. 5. Nodal primordium rate (PR) and the second phase of nodal leaf rate (LR2) versus mean temperature. Data from the greenhouse and field experiments were pooled and each line represents the fitted linear regression model. Circled data points were excluded from the PR regression analysis (for justification, see section on "Temperature influence on nodal primordium and nodal leaf rates" in Results section). See Figs. 2 and 3 for symbol code legends, and Table 1 for treatment descriptions. Inset: first phase of nodal leaf rate (LR1) versus mean temperature.

(r2 > 0.92; P < 0.001), despite the diversity in weather, soil, and management across treatments (Fig. 5). The fitted regression model indicated that LR2 increased from 0.09 leaves d-1 (i.e., 11 d leaf-1) at 15 °C to 0.26 leaves d-1 (i.e., 3.8 d leaf-1) at 25 °C (Fig. 5). Values of PR increased from 0.18 primordia d-1 (i.e., 5.6 d primordium -1) to 0.46 primordia d-1 (i.e., 2.2 d primordium-1) along the same 15-25 °C temperature range (Fig. 5). The plastochron and phyllochron parameters obtained from the inverse of the PR and LR2, were 36 and 58 °Cd, respectively. There was also a significant relationship between LR1 and air temperature during the NL1 to X1 timeframe (r2 = 0.64; P < 0.001) (Fig. 5, inset), which translated into a temperature-based phyllochron

of 83 °Cd and Tb of 9.5 °C for the time period between NL1 and Xj.

The relationship between PR or LR2 with temperature was not statistically different (P > 0.15) relative to their respective estimated Tb parameter values (9.6 °C versus 8.5 °C) (Fig. 5), and were also almost identical to the Tb for LR1 (9.5 °C). Note, however, that the slope of PR versus temperature was ca. 65% higher than for LR2 (i.e., 0.028 versus 0.017 laves d-1 °C-1), indicating that primordium formation was much faster than leaf appearance, even though the air temperature at the SAM was the same for both processes (Fig. 5). The estimated phyllochron for the first phase of leaf appearance (58 °Cd) represented ca. 70% of the phyllochron estimated for the second phase (83 °Cd). In other words, when expressed on a thermal time basis, LR2 was ca. 30% faster than LR1 (i.e., rate before the X1 breakpoint). The plastochron-to-phyllochron ratio changed from 0.4 before to 0.6 after the breakpoint. The pattern of primordium formation and leaf appearance in the SD expt., plotted on a thermal time scale, were almost identical across sowing dates, despite differences in mean photoperiod among treatments and also during the growing season (Supplementary material, Fig. S1). This observation indicated that there was little or no photo-period effect on plastochron and phyllochron, which was consistent with previous studies (Nico et al., 2015; Thomas and Raper, 1983).

3.4. Co-ordination between number of nodal primordia and nodal leaves

There was a strong association between NP and NL (R2 = 0.98; P < 0.001), which was stable across experiments, temperature regimes, and cultivars (Fig. 6). This close association indicates that, though PR is faster than LR, these processes are coordinated (Fig. 6). The model was also robust in terms of predicting the NP and NL reported 55 years ago by Johnson et al. (1960), who examined determinate and indeterminate soybean cultivars. Based on our fitted model, by the time of appearance of NL1 (cotyledons), soybean plants had already formed a total of five primordia, corresponding to the cotyledons, unifoliolate, and three trifoliolates (Fig. 4C-F). The ratio between appeared leaves and formed primordia abruptly changed from 0.4 to 0.7 appeared leaves per formed primordium when soybean plants had ca. 13 primordia and four leaves, which fit within the 2-6 range of leaves present at the X1 values estimated for the various experiments. The NPMAX was

Fig. 6. Relationship between number of nodal primordia (NP) and number of nodal leaves (Nl) during plant ontogeny, based on the pooled data from the greenhouse and field experiments. See Figs. 2 and 3 for symbol code legends and Table 1 for treatment descriptions. A two-segment linear regression model was fitted following Eqs. (5) and (6). Inset: relationship between the difference in maximum number of nodal primordia (Npmax) and nodal leaves (NLMAX) and cumulative thermal time from R3 to R5 (beginning of pod setting and seed filling, respectively; Fehr and Caviness, 1977). Thermal time (°Cd) was calculated as the sum of half-hourly measured temperature minus a base temperature (Tb = 9 °C), with the latter derived from the relationships between nodal leaf rates and temperature shown in Fig. 5. Data on NL collected after reaching the NPMAX are not shown here and were not used for fitting the two-segment model. Data from Johnson et al. (1960) were also plotted for comparison (stars).

consistently higher than the NLMAX. Clearly, this indicated that some of the primordia formed just prior to R3 did not advance to become fully-developed leaves at the stem apex before R5 stage (Fig. 6; Supplementary material, Tables S4 and S5). The difference between NPMAX versus NLMAX that we observed at end-season was actually inversely related to the length of just the R3-R5 phase of the season, when that phase was expressed in thermal time (r2 = 0.46; P < 0.001) (Fig. 6, inset). That phase spanned the timeframe between the end of pri-mordium formation at R3 and the (later) end of leaf appearance at R5 (Figs. 2 and 3). Hence, a longer duration of the R3-R5 phase, in terms of thermal time units, would seemingly allow a greater proportion of primordia formed after R3 to eventually become leaves (Supplementary material, Fig. S2).

4. Discussion

This is the first study to document the influence of temperature on the seasonal dynamics and biological synchrony of the formation of nodal primordia and appearance of nodal leaves in soybean. The experiments were conducted in both field and GH settings using indeterminate soybean cultivars differing in maturity, using frequent seasonal observations and sampling over the course of the growing season. Our results indicated that both PR and LR were temperature-sensitive. Moreover, the temperature-based phyllochron during crop ontogeny had two phases (i.e., changed from 83 to 58 °Cd), with an average breakpoint at NL4, whereas the temperature-based plastochron had a single phase (36 °Cd). However, both processes exhibited an almost identical Tb (ca. 9 °C) for the MG 2.1 and 3.0 indeterminate cul-tivars. The phyllochron reported here for the second phase of leaf appearance (58 °Cd) was almost identical to the phyllochron (56 °Cd) reported by Hesketh et al. (1973) and Thomas and Raper (1976), although there is a 3-degree difference in Tb between these two studies

(9 °C versus 6 °C). Estimated Tb in the prior two studies should be treated with caution since air temperature at SAM height was not likely measured. The findings of the present study do not support the findings made by Bastidas et al. (2008) about the lack of temperature-sensitivity of Lr amongst four sowing dates. The discrepancy is probably due to the very narrow range of mean temperature during the period of leaf appearance (22-26 °C) explored across treatments in their study.

We extended the knowledge on soybean primordium dynamics by documenting a definitive linear pattern in primordium formation that began at sowing and clearly ceased near R3 stage. In previous studies, PR was reported on a day-scale basis to be 0.50 to 0.52 primordia d-1 (i.e., 1.9- 2.0 d primordium-1) in determinate soybean cultivars grown at a mean temperature ranging from 23 to 24 °C in controlled environments (Chiera et al., 2002; Thomas and Kanchanapoom, 1991). This PR range is higher than the PR we observed (0.40 to 0.43 primordia d-1), based on the linear model fitted to our data for the same temperature range (Fig. 5). Possible explanations for this difference include (i) PR from these previous studies were based on SAM samples collected only in early vegetative development (i.e., only 10 days after VE or 28 DAS), compared to our study (i.e., 40 to 85 DAS); (ii) we recorded temperatures at SAM height, which was not done in prior studies; (iii) we recorded soil temperatures from sowing to VE of the seedling stem tip, which was not done in prior studies; and (iv) determinate cultivars were used in prior studies whereas we used indeterminate cultivars.

Our data clearly indicated that primordium formation completely ceased near stage R3. However, leaf appearance continued until R5 in indeterminate soybean in our study, as it also did in Bastidas et al. (2008) study. These two stages coincided with beginning of pod setting and seed filling, respectively. The cessation of leaf appearance at R5 in indeterminate soybean has been hypothesized to be due to photo-synthate diversion from the SAM to developing seeds by several authors (Bastidas et al., 2008; Egli et al., 1985; Pedersen and Lauer, 2004; Sinclair, 1984a). Relative to the main SAM in the two indeterminate cultivars, we were unable to conclusively observe the development of a terminal inflorescence meristem (Fig. 4G, H). This finding is consistent with current understanding of the genetic control of indeterminate versus determinate plants (Benlloch et al., 2015; Liu et al., 2010; Shannon and Meeks-Wagner, 1991). We do note here that Caffaro et al. (1988) reported observing a terminal inflorescence in indeterminate soybean. This apparent discrepancy can be explained by Carlson and Lersten (2004), who reported that a terminal SAM may sometimes appear to be a terminal floral inflorescence, but in reality, such an inflorescence is a series of small one- or two-flowered axillary inflorescences crowded together because of the short internodes near the main stem tip. Such an explanation is consistent with observations of differentiated floral primordia in the axillary apex adjacent to the main terminal SAM in the present study (Fig. 4H).

An original finding from this study is the temporal timing contrast between PR and LR. While plastochron was constant from sowing to end of primordium formation (i.e., near R3 stage), the phyllochron decreased from 83 °Cd (first phase of leaf appearance) to 58 °Cd (second phase of leaf appearance). Hesketh et al. (1973) and Fehr and Caviness (1977) also noted that LR was lower during the early vegetative development. Indeed, the phyllochron we derived in this study for the first phase of leaf appearance was not much different from the phyllochron of 106 °Cd that we derived from our re-analysis of Hesketh et al. (1973) data. The X1 breakpoint could not be fully associated with just a specific NL in the two to six range, nor could it be associated with the transition from heterotrophic (dependence on cotyledonary reserves) to auto-trophic (photosynthesis-driven) phases as reported in other crop species (Miralles et al., 2001; Padilla and Otegui, 2005; Sadras and Villalobos, 1993). More research is needed to better understand the mechanisms or environmental conditions that govern the timing of the first breakpoint (X1) in the soybean leaf appearance pattern.

Another novel finding from this study is that in indeterminate soybean, the NPMAX was always in excess of NLMAX. The magnitude of

this 'surplus' seemed to depend upon the length of the R3-R5 stage period expressed in thermal time. The dates of the R3 and R5 stages matched the dates when the accrual of nodal primordia and leaves ceased, respectively. It was evident in the data obtained in our study that increasing thermal time between R3 and R5 stages was associated with a decrease in the difference between NPMAX and NLMAX. As noted by Bastidas et al. (2008), Board and Tan (1995), and Board et al. (1999), the number of nodes per main stem can influence yield determination in soybean. Therefore, extension of the duration of the R3-R5 phase (in thermal time) may be a route to maximize NLMAX by R5 stage, which, in turn, can contribute to increase seed number and seed yield. Consistent with this hypothesis, Kantolic and Slafer (2001) and Nico et al. (2015) reported that an increase in the time period between R3 and R5 through artificial photoperiod extension resulted in greater node number, seed number, and seed yield.

Relative to the coordinated seasonal correspondence between pri-mordium formation and leaf appearance, our results showed that a bi-phasic linear model explained 98% of the PR and LR co-variation. Clearly, there was a strong synchrony between formation of primordia and appearance of leaves in indeterminate soybean, which was consistent with the relationships reported for other crop species (Miralles et al., 2001; Nemoto et al., 1995; Padilla and Otegui, 2005; Sadras and Villalobos, 1993; Turc and Lecoeur, 1997). In soybean, however, the coordination model had two linear phases of differing slopes that reflected a change in the plastochron-to-phyllochron ratio before and after an early season breakpoint (X1). The co-ordination model was stable across experiments, temperature regimes, and cultivars and was consistent with previous data on leaf appearance and primordia reported for soybean in the literature. Hence, the regression equation computed for the co-ordination model provides a convenient means of predicting NP existing at any given accrued number of visible appeared leaves, without the need for SAM dissection, which is very laborious and time consuming. The co-ordination model presented here can also be embedded into crop simulation models for robust simulation of seasonal primordium and leaf dynamics in indeterminate soybean.

5. Conclusions

Our study indicated that both PR and LR depended on temperature. While the response of PR to temperature was consistent during the entire crop cycle, LR response changed with ontogeny. Nodal pri-mordium formation and leaf appearance leaves ceased near the dates of R3 and R5 stages, respectively. There was a strong co-ordination between primordia formation and leaf appearance, which was stable across genotypes and temperature treatments.

Acknowledgments

We thank Nicolas Cafaro and Dr Alencar Zanon for their help in the field experiments. We also thank Dr. Ellen Paparozzi and Elizabeth Conley (University of Nebraska-Lincoln) for giving access to their laboratory, and to Michael Livingston, Aaron Hoagland, and Jeffrey Witkowski for their technical assistance.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.10167j.fcr.2017.03.015.

References

Bastidas, A.M., Setiyono, T.D., Dobermann, A., Cassman, K.G., Elmore, R.W., Graef, G.L.,

Specht, J.E., 2008. Soybean sowing date: The vegetative, reproductive, and agronomic impacts. Crop Sci. 48, 727-740. Benlloch, R., Berbel, A., Ali, L., Gohari, G., Millan, T., Madueno, F., 2015. Genetic control

of inflorescence architecture in legumes. Front. Plant Sci. 6, 1-14. Board, J.E., Tan, Q., 1995. Assimilatory capacity effects on soybean yield components and

pod number. Crop Sci. 35, 846-851.

Board, J.E., Kang, M., Harville, B.G., 1999. Path analysis of the yield formation process for late-planted soybean. Agronomy J. 91, 128-135.

Board, J.E., 2004. Soybean cultivar differences on light interception and leaf area index during seed filling. Agronomy J. 96, 305-310.

Caffaro, S.V., Martignone, R.A., Torres, R., Nakayama, F., 1988. Photoperiod regulation of vegetative growth and meristem behaviour toward flower initiation of an indeterminate soybean. Bot. Gazet. 149, 311-316.

Caffaro, S.V., Nakayama, F., 1988. Vegetative activity of the main stem terminal bud under photoperiod and flower removal treatments in soybean. Aust. J. Plant Physiol. 115, 159-166.

Carlson, J.B., 1973. Morphology. In: Caldwell, B.E. (Ed.), Soybean: Improvement, Production and Uses. American Society of Agronomy, Madison, pp. 17-66.

Carlson, J.B., Lersten, N.R., 2004. Reproductive morphology. In: Boerma, H.R., Specht, J.E. (Eds.), Soybeans: Improvement, production, and uses, 3rd ed. ASA, Madison, WI, pp. 59-95.

Chiera, J., Thomas, J., Rufty, T., 2002. Leaf initiation and development in soybean under phosphorus stress. J. Exp. Bot. 53, 473-481.

Egli, D., Guffy, R.D., Legget, J.E., 1985. Partitioning of assimilate between vegetative and reproductive growth in soybean. Agronomy J. 82, 69-73.

Egli, D.B., Zhen-wen, Y., 1991. Crop growth rate and seeds per unit area in soybean. Crop Sci. 31, 439-442.

Fehr, W.R., Caviness, C.E., 1977. Stages of soybean development. Special Report 80. Iowa State University, Cooperative Extension Service, Ames.

Granier, C., Tardieu, F., 1998. Is thermal time adequate for expressing the effects of temperature on sunflower leaf development? Plant Cell Environ.V 21 695-703.

Hay, R.K.M., Kirby, E.M., 1991. Convergence and synchrony- a review of the coordination of development in wheat. Aust. J. Agri. Res. 42, 661-700.

Hesketh, J.D., Myhre, D.L., Willey, C.R., 1973. Temperature control of time intervals between vegetative and reproductive events in soybeans. Crop Sci. 13, 250-254.

Johnson, H.W., Borthwick, H., Leffel, R.C., 1960. Effects of photoperiod and time of planting on rates of development of the soybean in various stages of the life cycle. Bot. Gazet. 122, 77-95.

Kantolic, A.G., Slafer, G.A., 2001. Photoperiod sensitivity after flowering and seed number determination in indeterminate soybean cultivars. Field Crops Res. 72, 09-118.

Kiniry, J.R., Rosenthal, W.D., Jackson, B.S., Hoogenboom, G., 1991. Predicting leaf development of crop plants. In: Hodges, T. (Ed.), Predicting crop phenology. CRC Press, Boca Raton, Florida, pp. 29-42.

Kirby, E.J.M., 1990. Coordination of leaf emergence and leaf andspikelet primordium initiation in wheat. Field Crops Res. 25, 253-264.

Lersten, N.R., Carlson, J.B., 2004. Vegetative morphology. In: Boerma, H.R., Specht, J.E. (Eds.), Soybeans: Improvement, production, and uses, 3rd ed. ASA, Madison, WI, pp. 15-57.

Liu, B.H., Watanabe, S., Uchiyama, T., et al., 2010. The soybean stem growth habit gene Dt1 is an ortholog of Arabidopsis TERMINAL FLOWER1. Plant Physiol. 153, 198-210.

Miksche, J.P., 1961. Developmental vegetative morphology of glycine max. Agronomy J. 53, 121-128.

Miralles, D.J., Ferro, B.C., Slafer, G., 2001. Developmental responses to sowing date in wheat, barley and rapeseed. Field Crops Res. 71, 211-223.

Nemoto, K., Morita, S., Baba, T., 1995. Shoot and root development in rice related to the phyllochron. Crop Sci. 35, 24-29.

Nico, M., Miralles, D.J., Kantolic, A.G., 2015. Post-flowering photoperiod and radiation interaction in soybean yield determination: Direct and indirect photoperiodic effects. Field Crops Res. 176, 45-55.

Padilla, J.M., Otegui, M.E., 2005. Co-ordination between leaf initiation and leaf appearance in f field-grown maize (Zea mays): Genotypic differences in response of rates to temperature. Ann. Bot. 96, 997-1007.

Pedersen, P., Lauer, J., 2004. Soybean growth and development in various management systems and planting dates. Crop Sci. 44, 508-515.

Rawson, H.M., 1993. Radiation effects on development rate in a spring wheat grown in different photoperiods and high and low temperatures. Aust. J. Plant Physiol. 20, 719-727.

Rawson, H.M., Zajac, M., 1993. Effects of higher temperatures, photoperiod, and seed vernalisation on development in two spring wheats. Aust. J. Plant Physiol. 20, 211-222.

Ritchie, J.T., NeSmith, D.S., 1991. Temperature and crop development. In: Hanks, J., Ritchie, J.T. (Eds.), Modelling plants and soil system, Agronomy series 31. ASA-CSSA-SSSA, Madison WI, pp. 5-29.

Sadras, V.O., Villalobos, F.J., 1993. Floral initiation, leaf initiation and leaf appearance in sunflower. Field Crops Res. 33, 449-457.

Saitoh, K., Nishimura, K., Kuroda, T., 2004. Characteristics of flowering and pod set in wild and cultivated types of soybean. Plant Prod. Sci. 7, 172-177.

Setiyono, T.D., Weiss, A., Specht, J., Bastidas, A.M., Cassman, K.G., Dobermann, A., 2007. Understanding and modeling the effect of temperature and daylength on soybean phenology under high-yield conditions. Field Crops Res. 100, 257-271.

Setiyono, T.D., Bastidas, A.M., Cassman, K.G., Weiss, A., Dobermann, A., Specht, J.E., 2011. Nodal leaf area distribution in soybean plants grown in high yield environments. Agronomy J. 103, 1198-1205.

Shannon, S., Meeks-Wagner, D.R., 1991. A mutation in the Arabidopsis TFL1 gene affects inflorescence meristem development. Plant Cell 3, 877-892.

Sinclair, T.R., 1984a. Cessation of leaf emergence in indeterminate soybeans. Crop Sci. 24, 483-486.

Sinclair, T.R., 1984b. Leaf area development in field-grown soybeans. Agronomy J. 76, 141-146.

Sinclair, T.R., Marrou, H., Soltani, A., Valdez, V., Chandolu, K.C., 2013. Soybean

production potential in Africa. Global Food Security 3, 31-40.

Slafer, G.A., 1995. Wheat development as affected by radiation at two temperatures. J. Agronomy Crop Sci. 175, 249-263.

Slafer, G.A., Rawson, H.M., 1994. Sensitivity of wheat phasic development to major environmental factors- a reexamination of some assumptions made by physiologists and modelers. Aust. J. Plant Physiol. 21, 393-426.

Slafer, G.A., Rawson, H.M., 1997. Phyllochron in wheat as affected by photoperiod under two temperature regimes. Aust. J. Plant Physiol. 24, 151-158.

Specht, J.E., Diers, B.W., Nelson, R.L., Toledo, J.F., Torrion, J.A., Grassini, P., 2014. Soybean (Glycine max (L.) Merr). In: Smith, J.S.C., Carver, B., Diers, B.W., Specht, J.E. (Eds.), Yield gains in major US field crops: contributing factors and future prospects. ASA-CSSA, Madison, WI.

Sun, C.N., 1957. Histogenesis of the leaf and structure of the shoot apex in glycine max (L.) merrill. Bull. Torrey Botanical Club 84, 163-174.

Thomas, J.F., Kanchanapoom, M.L., 1991. Shoot meristem activity during floral transition in glycine-max (L) merr. Bot. Gazet. 152, 139-147.

Thomas, J.F., Raper, C.D., 1976. Photoperiodic control of seed filling for soybeans. Crop Sci. 16, 667-672.

Thomas, J., Raper Jr., C.D., 1983. Photoperiod effects on soybean growth during the onset of reproductive development under various temperature regimes. Bot. Gazet. 144, 471-476.

Turc, O., Lecoeur, J., 1997. Leaf primordium initiation and expanded leaf production are co- ordinated through similar response to air temperature in pea (Pisum sativum L). Ann. Bot. 80, 265-273.