| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Lousiana State Univ. Agric. Center, P.O. Box 438, Saint Joseph, LA 71366
b Texas A&M Univ., Dep. of Soil and Crop Sciences, College Station, TX 77843-2474
c Lousiana State Univ., Dep. of Experimental Statistics, 161 Agric. Administration Building, Baton Rouge, LA 70803-5606
d 1355 Commerce Dr., Auburn, AL 36830. Partial support of this research was provided by the Cotton Foundation
* Corresponding author (eclawson{at}agcenter.lsu.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: BMP, boll maturation period • CR, conventional row • DAP, days after planting • DD15, growing degree days with base temperature 15.6°C • N7P1, main stem node 7, sympodial branch position 1 • UNR, ultra narrow row
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
1 Throughout the current manuscript, the term "boll set" is applicable only to bolls retained through harvest. 
Received for publication April 10, 2007.
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In previous research, the occurrence of early maturity in UNR relative to CR systems has been somewhat inconsistent. Jost and Cothren (2001) found evidence of substantially earlier maturity for UNR than for CR treatments in the initial study year, but did not find treatment differences in the following year. Galanopoulou-Sendouka et al. (1980) found that UNR treatments were significantly earlier in mean maturity date than a CR treatment within an early planting date in one study year. However, this was not the case in other combinations of year and planting date. Baker (1976) found no significant differences in percentage first harvest between high population-narrow row production systems and CR cotton. Most reports have shown evidence of either similar or earlier timing of maturity for high population-narrow row systems as compared to CR cotton. However, the opposite effect is suggested by the results of Fowler and Ray (1977) who reported, for a 1-yr study, several instances of later mean maturity date with increased plant population in equidistant plant spacing. There is a need to more fully establish the expected relationships of UNR and CR cotton in the timing of maturity and to examine the underlying causes of such relationships.
Ultra-narrow row systems differ from CR systems in both row spacing and plant population, and the effects of these factors cannot be distinguished in studies in which they were not varied independently. However, the theory that earlier maturity in UNR cotton is derived from the use of elevated plant populations is in many cases supported by the responses of cotton to elevated plant population in the literature. Both Smith et al. (1979) and Mert et al. (2006) found that in some or all cases, a higher percentage of the yield was obtained in the first harvest with higher plant population, indicating earlier maturity. Jones and Wells (1997) found patterns of boll opening indicative of earlier maturity in normal relative to low plant population. As plant populations were increased, Bednarz et al. (2000) found fewer bolls plant–1 and, without differences in total bolls m–2, greater numbers of bolls m–2 in the first sympodial position. These effects can be conducive to earlier maturity. In contrast, Kerby et al. (1990) reported that the timing of maturity was unaffected or delayed by higher plant population, depending on the growth habit of the genotype. Mohamad et al. (1982) found evidence of delayed maturity at higher plant populations. As is the case for comparisons of UNR and CR cotton, the evidence suggests that increases in plant population are conducive to earlier maturity, but the effect has not been consistent.
Although the hypothesis has not been extensively evaluated, it is not implausible for row spacings to affect the timing of maturity independently of plant population. At plant populations typical of commercial cotton production, narrow row spacings provide more uniform interplant spacing. Therefore, as noted by Buxton et al. (1979), narrow rows potentially allow better use of early season resources. Heitholt et al. (1993) compared 0.5- and 1.0-m row spacings in normal and okra-leaf isolines of cotton without variation in plants m–2. Earlier attainment of 65% of lint on open bolls was observed in the narrower row spacing within the okra-leaf isoline in one of three environments. No other significant row spacing effects on the timing of maturity were observed. Hawkins and Peacock (1973) report a lack of significance for the effects of row spacing, plant population, and the interaction of row spacing and plant population on percentage first harvest. Neither study suggests strong influences of row spacing, when independent of plant population, on the timing of maturity. However, additional research may be needed to confirm the general applicability of the results.
Nitrogen fertilizer, an important input in most cotton production systems, can influence the timing of maturity. Rochester et al. (2001) found that delays in the attainment of 60% open boll varied by experiment, but ranged from 0.6 to 3.2 d per additional increment of 50 kg N fertilizer ha–1. In work reported by Constable et al. (1992), additional increments of 100 kg N fertilizer ha–1 resulted in 5- to 6-d delays in the estimated date of 60% harvest. Boman and Westerman (1994) reported that the percentage first harvest was reduced by up to 17.4 percentage points by higher N rates in the first study year. Although responses differed in the following year, these results were influenced by rainy conditions that interfered with boll opening after the first harvest. The results of MacKenzie and van Schaik (1963) included some reductions in percentage first harvest with greater N. In Pima cotton (G. barbadense L.), Tewolde et al. (1994) found instances in which the percentage open bolls was higher for cotton receiving no N than for cotton to which N was applied. If early maturity is an objective, effects such as these may have implications for UNR N fertilization practices, and any interaction of N rate and row spacing (i.e., UNR vs. CR cotton) on the timing of cotton maturity may be important to the selection of appropriate N rates for UNR cotton. In addition, such an interaction could have a bearing on the relative merits of UNR and CR production systems in a given environment, such as one fertilized for a high yield goal.
The morphological and developmental characteristics affected by UNR systems and N rates are important means of understanding their effects on cotton crop maturation. A reduction in bolls plant–1 was found in most instances for UNR relative to CR cotton by Jost and Cothren (2000), and has also been found in some cases due to reductions in N fertilizer rate in Upland (Wiatrak et al., 2005) and in Pima cotton (Tewolde et al., 1994). Cotton plants in UNR and CR systems can also differ in the physical distribution of bolls over ranges of main stem nodes and sympodial branch positions (Jost and Cothren, 2000; 2001). Boquet et al. (1993) showed variation, with increasing N, in the probability of occurrence of harvestable bolls at specific sympodial positions within ranges of main stem nodes. The differences between UNR and CR systems reported by Galanopoulou-Sendouka et al. (1980) included a higher first fruiting node in 1 UNR spacing (1 of 2 yr), earlier beginning of flowering in both UNR spacings (1 combination of year and planting date only), a longer boll maturation period (BMP) in 1 UNR spacing (1 yr only) and a higher value for a ground-area based index of flowering rate in both UNR spacings. Several other characteristics can also be related to the timing of maturity. Examples, with important related papers, include the node of the first sympodial branch (Ray and Richmond, 1966), vertical and horizontal flowering intervals (intervals between successive flowering events for bolls identical in sympodial branch position on successive main stem nodes and for bolls in successive positions on the same sympodial branch, respectively) (Munro, 1971), ground area-based rates of flower retention and boll set1 (Jones and Wells, 1997), and the timing of cutout as indicated by the occurrence of five nodes above white flower (Bourland et al., 2001).
Many plant and crop characteristics that are related to the timing of maturity describe some aspect of the temporal distribution of boll set. The node of the first sympodial branch and the timing of cutout, for example, provide insights into the beginning and ending points of this distribution. Vertical and horizontal flowering intervals and rates of boll set or blooming may delineate periods of time in which there was potential for boll set to be more concentrated. Because fruiting sites are created sequentially, the distribution of bolls across fruiting sites can provide insights into the distribution of boll set over time. However, there are at least two characteristics that can influence the timing of maturity without changes in the temporal distribution of boll set. One is BMP, both as a season-long average and by variation with date of bloom. A second, if the timing of maturity is measured in terms of percentages of seedcotton weight harvested, is temporal variation in boll weight. Examination of indicators both related and unrelated to the temporal distribution of boll set contributes to a comprehensive understanding of the influence of treatments on the timing of maturity. This may provide insights into management pratices needed to maximize early maturity and conditions under which effects on the timing of maturity can be expected.
The objectives of this report are to characterize temporal aspects of UNR and CR cotton crop maturation and, using this information along with other plant and crop characteristics, to identify mechanisms by which the timing of maturity was affected in each system. A similar analysis is presented for a series of N fertilizer rates. Interactions of N rate and row spacing, where present, are examined.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The experimental design was a split plot. Nitrogen fertilizer rates of 0, 50, 101, and 151 kg ha–1 were imposed as whole plots. These were equivalent, respectively, to 0.0, 0.5, 1.0, and 1.5 times the 2000 N fertilizer recommendation provided by the Texas A&M University System Soil Testing Laboratory, College Station, TX. The N fertilizer was urea ammonium nitrate (32–0–0 N–P–K), surface applied and incorporated within 4 h. The split plots were row spacings of 19, 38, and 76 cm. These were established using a Great Plains grain drill, blocking appropriate seed drop tubes to achieve the wider row spacings. The variety was ‘Deltapine 422 BG/RR.’ All row spacings were planted on a flat seed bed at high populations and hand thinned. Target plant populations were increased with each reduction in row spacing but held constant across N rates. Averaged over the 3 yr, plant populations were 296,600, 195,400, and 136,300 plants ha–1 in the 19, 38, and 76-cm row spacings, respectively, and fell within a range of 204,600 to 214,200 plants ha–1 for each N rate as averaged over row spacings (Clawson et al., 2006). Because the plant populations differed in each row spacing, the effects of plant population and row spacing cannot be distinguished. The test was replanted in both 2000 and 2001 because heavy precipitation caused soil crusting and stand failure following the initial planting. In 2000, fertilizer treatments were applied on 27 April. The field was initially planted on 28 April and was replanted on 15 May. In 2001, fertilizer treatments were applied on 19 April. The field was initially planted on 4 May and was replanted on 18 May. In 2002, fertilizer treatments were applied on 24 April and a satisfactory stand was achieved from cotton planted on 1 May.
Mepiquat chloride (N, N-dimethylpiperidinium chloride) was applied uniformly across all treatments. Applications of 24.5 g a.i. ha–1 were made on 16 June 2000, 21 June and 4 July 2001, and 12 June and 12 July 2002. Before the final application in 2002, an additional application of mepiquat chloride was washed off by rain. Insect pest pressure was greatest in 2001 and least in 2002. The fields were irrigated using an overhead linear-move sprinkler system. Climatic conditions were variable among the study years. During 2000, high temperatures were experienced in the latter part of the growing season. In 2001, frequent rain occurred during harvest. The year 2002 was characterized by moderate temperatures and plentiful, well distributed rainfall. Additional information on climate, production practices, plant populations, and soil properties is reported in Clawson et al. (2006).
Data Collection and Analysis
All analyses of variance were performed using PROC MIXED of SAS at the 0.05 level of significance (SAS Institute, 2002). The results are reported as least squares means separated by the Tukey–Kramer method. Data were combined over years, with years considered to be a random effect. Because of weather-induced irregularities and other considerations, the years over which the data were combined varied by response variable (Table 1
). With a small number of exceptions for simplicity, comparability to other response variables, or both, analyses included the maximum number of years for which adequate data were available. Further description, including additional statistical procedures if appropriate, is provided below for the individual response variables.
|
|
Individual Boll-derived Measurements
The area designated for monitoring of individual bolls consisted of a 1-m length of row in each plot. This area represented 0.19, 0.38, and 0.76 m2 in the 19-, 38-, and 76-cm row spacings, respectively. Based on 3-yr average plant populations, which were measured elsewhere in the plots, it contained 5.6, 7.4, and 10.4 plants in the 19-, 38-, and 76-cm row spacings, respectively. During flowering, the designated area was checked at least every other day and a paper tag with the date of white flower occurrence was attached to each bloom. Unless very small, bolls were defined as open if the distance between the two most widely separated carpel tips was a minimum of approximately 1 cm. Bolls were harvested on the first day they were open. After the great majority had opened, the few remaining bolls were collected during a one-time harvest following an application of harvest aids. The white flower date, boll opening date (for bolls collected before the harvest aid application), and main stem node and sympodial branch position were recorded for each boll. Bolls located on monopodial branches were noted as such without further description of fruiting site. Following its removal and air drying, the weight of the seedcotton from each boll was recorded as well.
There were occasional instances of ambiguity between two adjacent dates of white flower or boll opening for reasons such as a missed day of bloom tagging or individual boll harvest. If the missing date was assigned to these bolls, it was done as an estimate midway between the two possible dates under the assumption that this would be correct for the bolls on average. Data were excluded from analyses for bolls from plants with terminal damage in 2000 and 2002. In 2001, data were included for such bolls if located on nodes below that of the terminal damage. Data were also included from 2001 for all bolls from plants on which a monopodial branch had clearly taken over as the main stem following terminal damage. In these cases, the junction of this branch with the original main stem was counted as two nodes under the assumption that bud break from the node of terminal damage took the time required to add one main stem node. Data were not collected on bolls with substantial damage or that were located on damaged branches or peduncles in 2001 or 2002. However, data were collected on such bolls in 2000 and these bolls were included in analyses that did not involve seedcotton weight or boll opening date. Bolls containing 1 g or less of seedcotton were considered to be unharvestable and with the exception of the date of main stem node 7, sympodial position 1 (N7P1) boll set were excluded from analyses.
Boll maturation period was defined as the number of days between the dates of white flower and boll opening for an individual boll. In a separate analysis, BMP was expressed in growing degree days with base temperature 15.6°C (DD15). The criteria for boll opening imply that the definition of BMP differs slightly from that of studies, such as Bednarz and Nichols (2005), in which cracked bolls were considered to be open. Data were combined only over 2000 and 2002. Bolls with estimated white flower or opening dates comprised 4.2 and 0.5% of the datasets for BMP in 2000 and 2002, respectively. Bolls with an unknown and unestimated date of white flower made up 21 and 10%, respectively, of the datasets that could otherwise have been used for BMP in 2000 and 2002. This resulted from the difficulty of locating all blooms in the designated areas of a 48-plot study as well as some losses of the paper tags that identified the date of white flower. It also reflected a slightly premature cessation of bloom tagging in 2000, as evidenced by a sharp increase, after 10 September, in the percentage of harvested bolls for which the date of white flower was unknown. However, more than 90% of the bolls that would have been usable in the analysis of BMP if date of white flower was known had been harvested by that time. Dates of white flower of bolls in the BMP dataset ranged from 3 to 28 July in 2000 and from 21 June to 28 July in 2002. Dates of boll opening in this dataset ranged from 13 August to 24 September in 2000 and from 5 August to 12 September in 2002.
An additional analysis was performed to assess the responses of BMP, expressed in days, to date of boll set. Linear regression of BMP on date of white flower was performed for each individual plot in 2000 and 2002 using PROC REG of SAS (SAS Institute, 2002). The slopes resulting from these regressions (BMP slope) were subjected to analysis of variance combined over 2000 and 2002. A visual assessment of the data for each combination of N, row spacing, and year provided little evidence of a quadratic component in these effects, and fitting equations of the form y = a + bx + cx2 did little to improve R2 values for these year and treatment combinations. Although these comparisons were not made for individual plots, it appears that a linear relationship adequately described the effects of date of boll set on BMP.
Temporal variation in boll weight was analyzed by linear regression of boll weight on date of white flower for each individual plot, using PROC REG of SAS (SAS Institute, 2002). For simplicity, this analysis was performed on the same set of bolls that was analyzed for the determination of BMP and BMP slope. The slopes resulting from these regressions (boll weight slope) were subjected to analysis of variance combined over 2000 and 2002. Bednarz et al. (2000) found that the weight of bolls in first, second, and third sympodial positions increased up to an optimal main stem node specific to that position, after which it declined. Similar patterns were present in the results of Jenkins et al. (1990). This indicates the possibility of a quadratic component in the response of boll weight to date of boll set, although comparisons to the current analysis, which integrates all fruiting sites, are somewhat problematic. The current study data contained few visual indications of quadratic responses in any combination of N, row spacing, and year, and fitting equations of the form y = a + bx + cx2 did little to improve R2 values for these year and treatment combinations. While this assessment was not made for individual plots, it appears that the linear relationship adequately represented the responses of boll weight to date of boll set.
Boll set rate was defined as the number of bolls set m–2 (5-d flowering period)–2. It accounts only for bolls retained through harvest. Data were combined over 2000 and 2002. The first day of the initial flowering period in each year was the earliest date of white flower for a harvested boll in the entire study. Averaged over 2000 and 2002, flowering periods 1, 2, 3, 4, and 5 represented 50–54, 55–59, 60–64, 65–69, and 70–74 DAP, respectively. In each of these years, the majority of bolls were set during the five flowering periods analyzed. An assessment of bolls by picking date indicated that averaged over 2000 and 2002, date of white flower was probably missing from approximately 13% of the bolls originating during these flowering periods, and that missing values were probably more common during the later portion of the 25-d interval represented. This implies some underestimation of boll set rate, particularly during the later flowering periods. However, there was little evidence of treatment effects on the percentage of bolls for which the date of white flower was unknown (data not shown), suggesting that the dataset was satisfactory for comparisons of boll set rate between treatments.
The date of first boll set was defined as the earliest date of white flower for a harvested boll from each plot. Data were combined over 2000 and 2002. Results were expressed in DAP.
The date of N7P1 boll set was defined as the average date of white flower for bolls harvested from the first sympodial position at main stem node 7. A similar parameter was defined by Munro (1971). The flowering dates for bolls with 1 g or less of seedcotton were included in this analysis. Data were combined over 2000, 2001, and 2002. Results were expressed in DAP.
Physical Boll Distribution and Bolls Per Plant
In 2001 and 2002, physical boll distribution was mapped for each plot. For this purpose, six representative plants from each plot were cut at random from areas not set aside for other purposes. The location of each boll was noted in terms of main stem node (counting the cotyledonary nodes as zero) and sympodial branch position (counting the site proximal to the main stem as position 1). The bolls originating on monopodial branches were noted as such without further description of fruiting site. Counts were made of total bolls, bolls in specific sympodial positions and ranges of main stem nodes, and bolls on monopodial branches. In 2000, physical boll distribution data were compiled from the locations of the bolls in the area designated for individual boll measurements. Physical boll distribution data were subjected to analyses of variance combined over 2000, 2001, and 2002.
The numbers of bolls plant–1 were obtained from the six plants that were mapped for physical boll distribution in 2001 and 2002.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
|
|
|
The mechanisms causing the differences among row spacings in the patterns of boll set rates appear to include both physical boll distribution and boll load. Jones and Wells (1997) found delays in the attainment of high rates of flowering in a treatment with greater proportions of total bolls on monopodial branches. They attributed this to the time required to develop sufficient plant architecture to support flowering at such fruiting sites. Although their study examined plant populations only, the principle that was advanced appears to be generally applicable to practices that increases the proportion of bolls on fruiting sites not proximal to the main stem. In the current study, wider row spacings were associated with lesser percentages of first position bolls and greater percentages, in general, of second position bolls and bolls on monopodial branches (Table 7 ). Because bolls on monopodial branches or in distal sympodial positions can be set concurrently with bolls more proximal to the main stem at higher main stem nodes, the physical boll distribution in wider row spacings was conducive to faster rates of individual-plant boll set. As noted by Jones and Wells (1997), this effect would be expected to be delayed by the time required to produce flowers outside of the first position. Continuing this line of reasoning, greater delays in attainment of rapid boll set would be expected from physical distributions of bolls that were more widely spread relative to the main stem. In the current study, differences in physical boll distribution were moderate in extent, and the great majority of bolls were located in the first and second positions in all row spacings. This finding is probably reflected in the small size of the differences among row spacings in the timing of harvest stages.
|
Because of the developmental pattern of the cotton plant, the influences of physical boll distribution and boll load on boll set rate probably differed in importance by flowering period. The second flowering period ended no later than 7 d after first boll set for any row spacing. Before that time, in all row spacings, it is likely that bolls were primarily set in the first position and that effects of boll loads on boll set rates were still minimal. This is compatible with the hypothesis, closely related to the discussion of Jones and Wells (1997), that a general lack of mechanisms and impetus for adjustment of individual-plant boll set rate to plant population during early bloom allowed faster boll set in the 19 cm relative to the 76 cm spacing (Fig. 1). Differences in physical boll distribution would have been more fully expressed after flowering period 2, and were conducive to faster individual-plant boll set in wider row spacings. Within the 38-cm spacing and especially the 76-cm spacing, this may have been an important means of increasing boll set rate m–2. Because of greater boll set rates during the second flowering period, effects of boll load may have become influential earlier or have been more pronounced with narrower rows, a factor conducive to slower rates of boll set m–2. A combination of these two effects may have prevented differences in boll set rates during flowering periods 3 to 5, despite substantial differences in plant population among the row spacings.
In the current study, the fact that differences among treatments, such as those of physical boll distribution and the putative accumulation of boll load, provided a potential means of counteracting plant population effects does not prove that they were caused by plant population. The row spacing treatments were production systems differing in both plant population and row spacing, and respective influences of these differences were not separable. Several reports have indicated that the effects of increasing plant population on physical boll distribution are similar in nature to those observed with reduction in row spacing in the current study (Jones and Wells, 1997; Bednarz et al., 2000; Kerby et al., 1990). In a study in which plant population was held constant on a ground area basis, Heitholt et al. (1993) did not show row spacing differences in percentages of bolls located within any of the main stem node ranges analyzed. However, this finding does not preclude differences in physical boll distribution by sympodial branch position, a variable not reported for that study. A more rapid increase in early boll load under high plant population is indicated by the Jones and Wells (1997) results. Buxton et al. (1979) found that relative to single rows at the same ground area-based population, two rows per bed resulted in lags in boll set during the first 2 wk of flowering. Although this suggests that the reduction in row spacing was not of itself conducive to faster accumulation of boll load, the authors noted that responses may have been influenced by the potential for greater herbicide effects with two rows per bed. Overall, the literature supports the idea that plant population contributed to the effects observed in the current study on physical boll distribution and putative accumulation of boll load. While not supportive per se, the literature reviewed did not contain strong evidence against influences on these characteristics from row spacing.
The results of this study have several implications for UNR cotton production. First, if boll set is delayed until substantial numbers of fruiting sites outside of the first position have developed in CR cotton, differences between the systems in initial rate of boll set m–2, initial rate of accumulation of boll load, and the timing of maturity may be reduced. For this reason, protection of early fruiting forms may be critical to the realization of earlier maturity in UNR relative to CR cotton. Second, it appears likely that differences in the timing of maturity between UNR and CR systems would be enhanced by conditions producing larger contrasts in the distribution of bolls to sympodial positions distal to the main stem (and in the degree to which these positions are distal) and to monopodial branches. Although their effects were not verifiable in the current study, differences in plant population have previously been shown to affect these aspects of physical boll distribution. For the 3 yr of the current study, the 19- and the 76-cm row spacings differed in plant population by a factor of approximately 2.2, but a valid comparison could also have been made with a lower population in the CR spacing. Further research comparing UNR and CR systems under greater contrasts in plant population may be justified. Finally, there were compensatory mechanisms that prevented the elevated rate of boll set in UNR cotton from extending beyond early stages of bloom. These mechanisms therefore limit the degree to which UNR is earlier in maturity than CR cotton. The value of 2 to 3 d earlier maturity, as was observed in the current study, is likely to be minimal when compared to the high seed costs and other expenses of a UNR production system.
Nitrogen Rate Effects
The N fertilizer rates used in this study had no effect on the timing of harvest stages (Table 3), suggesting a similar timing of maturity in each N rate when averaged over 2000 and 2002. This occurred in spite of increases in bolls plant–1 (Table 7) and, because of similar average plant populations, putative bolls ha–1 with greater N. Each incremental increase in N rate resulted in a significant increase in lint yield (combined over 2000 and 2002; data not shown), suggesting that no rate of N was excessive. In previous reports, the influence of differences in N fertilization within a nonexcessive range of rates on the timing of cotton maturity has been variable. As interpreted by the current authors, MacKenzie and van Schaik (1963) showed no significant reductions in percentage first harvest from higher N except within the set of comparisons that included an N rate above the rate that optimized yield. However, Constable et al. (1992) reported delays in 60% harvestable lint in association with positive yield responses to N, and in Pima cotton, Tewolde et al. (1994) found instances of a similar relationship. Certainly, greater yields are considered to be associated with later maturity. In the current study, it appears that N affected one or more aspects of cotton development in ways that prevented later maturity of the higher N rates despite their greater yields.
The average BMP was almost 1 d shorter in the 0 kg than in the 151 kg ha–1 N rate (Table 4), a pattern conducive to slightly earlier attainment of harvest stages in the 0 kg ha–1 N rate. When expressed in DD15, N rate effects on BMP were very similar. Given the differences in plant size among N rates, contrasts in the microclimate affecting bolls may have been responsible for the reduction in BMP in the 0 kg ha–1 rate of N. The values of BMP slope were near zero for all rates of N, with the slope of greatest magnitude, that of 0 kg ha–1, implying an approximately 1 d difference in BMP for bolls set 25 d apart. However, N effects were relatively close to significance (P > F = 0.1203; Table 1) and BMP slope was consistently decreased with lower N. If meaningful, and all else being equal, differences in BMP slope would have caused later attainment of 30% harvest, little effect on the timing of 60% harvest, and earlier attainment of 85% harvest with reductions in N rate. At the level of the observed differences in BMP slope these effects would have been small even if present.
Boll weight slopes were generally negative, indicating reductions in boll weight with later date of bloom (Table 5). Such reductions can affect the timing of harvest stages, but evidence of treatment differences in the reductions was limited. With the exception of comparisons with the 0 kg N ha–1, 19-cm treatment, boll weight slope did not differ among treatments differing in N rate. This appears to contrast with the results of Hearn (1975) who measured the weight of bolls from each of eight successive hand harvests. As interpreted by the current authors, reductions in boll weight during later harvests were generally larger and more consistent for cotton fertilized with 34 kg than with 112, 168, or 225 kg N ha–1. However, in the current study, average boll weight was reduced up to 12% in the 0 kg ha–1 relative to higher N rates (3-yr average; Clawson et al., 2006). For this reason, the general similarity in boll weight slope among N rates may imply a slightly more rapid decline in boll weight on a percentage basis in the 38- and 76-cm, 0 kg ha–1 treatments. If meaningful, and all else being equal, an identical boll weight slope in the presence of lower average boll weight would cause slightly earlier attainment of each harvest stage at 0 kg N ha–1. The effect would have been greater at 60% than at the other reported harvest stages, but under a 12% reduction in boll weight would have been extremely small in all cases.
In general, it does not appear that boll properties other than the date of boll set differed sufficiently to cause substantial differences among N rates in the timing of harvest stages. Boll maturation period was affected by N, but the magnitudes of the differences were small. Nitrogen rate effects on BMP slope and boll weight slope were not substantiated with a high degree of certainty. If present, it appears that each effect would have been minor, and in terms of implications on the date of 30% harvest, opposing in nature. However, the potential influence of each was assessed under assumptions of constant rates of boll set and in most cases, a lack of variation in other characteristics. If either was interactive with other types of variability in its effects, it is not impossible that the importance of BMP slope or BW slope to the timing of maturity could have been greater.
Rates of boll set were reduced with lower N (Fig. 2 ), and the interaction between N and flowering period on boll set rate was not significant (Table 1). The general reduction in boll set rate with lower N appear to have counteracted the potential effects of fewer bolls plant–1 (Table 7) on the timing of harvest stages.
|
Due to small size and chlorosis, plants in lower N rates were probably more limited in capacity to produce photosynthate. In theory, such source limitations could slow boll set rate by means that include reduced rates of fruit retention and longer flowering intervals. The data provide few insights into fruit retention rates. There were indications, however, of differences in rates of main stem node addition, a characteristic closely related to the vertical flowering interval. The most direct evidence for this is the delay in N7P1 boll set for the 0 kg ha–1 relative to the other N rates (Table 6). It is also suggested by data on total nodes and boll distribution. There were fewer total main stem nodes in the two lower N rates (3 yr average; Clawson et al., 2006). Likewise, the percentage of total bolls within the upper node range was generally reduced with lower N, a finding compatible with a hypothesis of termination of boll set at a lower main stem node. The similarity in timing of harvest stages suggests that the primary causes of these effects were slower rates of node and fruiting site addition with lower N. Boquet et al. (2004) reported that for cotton following a wheat cover crop or native winter vegetation, the number of main stem nodes at 55 DAP was reduced in 0 kg ha–1 relative to higher rates of N. In a greenhouse study, Jackson and Gerik (1990) found that reductions in main stem nodes on plants fertilized with lower N were evident as early as the approximate date of first bloom. These findings confirm that a reduction in rates of main stem node addition, beginning early in the season, can occur under conditions of N deficiency. Although flowering intervals were not assessed in the current study, it appears likely that there were extensions of the vertical flowering interval by means of slower node addition in lower N rates.
Relative to higher N rates, first boll set was delayed by 2 d in the 0 kg ha–1 N rate (Table 6). A similar effect was found for the date of N7P1 boll set. However, these delays did not translate into significantly later attainment of 30% harvest for the 0 kg ha–1 N rate (Table 3). During the flowering periods primarily contributing to 30% harvest, there is no indication that the boll set rates for the 0 kg ha–1 N rate were closer to those of higher N rates than they were thereafter (data not shown). Of the effects measured in this study, the slight reduction in BMP seems the most likely to be responsible, at least in part, for the lack of delay in 30% harvest despite later initiation of boll set in the 0 kg ha–1 N rate.
The literature contains examples in which suboptimal N rates were earlier in maturity than N rates producing greater yields (Constable et al., 1992; Tewolde et al., 1994), and at least one other in which they were not (MacKenzie and van Schaik, 1963). The results of the current study indicate that in some instances, a slower rate of boll set can prevent earlier maturity under low N rates. In light of the high probability of yield losses and the uncertain nature of any reduction in time to maturity, there is little justification for the application of suboptimal rates of N for the purposes of ensuring an early cotton harvest.
Interaction of Nitrogen Rate and Row Spacing
Significant interactions between N and row spacing were found only for boll weight slope and the timing of 30% harvest (Table 1). The interaction of N and row spacing on boll weight slope appears to have occurred mainly due to a high value for boll weight slope in the 0 kg N ha–1, 19-cm row spacing treatment (Table 5). The reasons for the higher boll weight slope in this particular treatment are not entirely clear, but may be related to the low numbers of bolls per plant in this treatment. The confinement of the unusual response to an unfertilized, N-deficient treatment suggests that the interaction has few implications for commercial cotton production.
The responses of the row spacings to N differed in two ways that may have contributed to the significant interaction between these factors in the timing of 30% harvest (Table 1). First, differences among row spacings were small within the 0 kg ha–1 in relation to those within the three higher N rates (Fig. 3 ). While the earliest and latest row spacing treatments differed significantly within each of the higher N rates, there were no significant differences among row spacings within the 0 kg ha–1 rate. Second, the 19-cm row spacing reached 30% harvest earlier (numerically) than the 38 cm only if fertilized with 101 or 151 kg N ha–1. Although the interaction was not significant (Table 1), the relationship between row spacings and N rates was similar in many respects for the timing of 60% harvest (data not shown). The reporting of main effects had a slight masking effect on the variability of crop maturation among treatments, but within individual N rates, no two row spacings varied by more than 3.7 d in the attainment of any harvest stage. The similarity between the 19- and 38-cm row spacings in the timing of 30 and 60% harvest (Table 3) appears to be a reflection of the inclusion of the data from the two lower N rates in the row spacing main effect means.
Low rates of N were generally not effective in enhancing the degree to which the UNR spacings were earlier than the CR spacing in 30% harvest (Fig. 3) or the other harvest stages (data not shown). Reductions in N fertilizer rates in UNR systems for this purpose are not justified by the results of the current study. Additional research is needed to verify a lack of interaction between N rate and row spacings at excessive rates of N, under which interplant competition and retention of early fruiting forms may play greater roles in the timing of maturity.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The limited extent of the differences among row spacings in the timing of maturity may be related to the moderate nature of the observed contrasts in physical boll distribution. It also reflects the mechanisms, probably related to physical boll distribution and boll load, which limited faster boll set with narrower row spacing to early bloom. It appears that retention of early fruiting forms is critical to achieving early maturity in UNR relative to CR cotton.
Nitrogen fertilizer rate did not affect the timing of 30, 60, or 85% harvest. The similarity in timing of maturity among N rates occurred despite reductions in lint yield, bolls plant–1, and putatively, bolls m–2 with lower N, as well as a slight reduction in BMP in the 0 kg ha–1 N rate. Earlier maturity with lower N was prevented by lower rates of boll set. Several types of data were indirectly indicative of reduced rates of main stem node addition with lower N. This suggests the possibility that longer vertical flowering intervals may have contributed to slower boll set with reductions in N. Unexamined characteristics, such as horizontal flowering intervals or rates of fruit retention, may also have contributed to slower boll set in lower N rates.
Interactions between N and row spacing, where present, did not imply a greater degree of early maturity of the UNR spacings relative to the CR with reductions in N rate. The results provide little justification in terms of the timing of maturity for reducing N fertilizer rates in UNR relative to CR cotton.
|
ACKNOWLEDGMENTS |
|---|
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
1 Throughout the current manuscript, the term "boll set" is applicable only to bolls retained through harvest.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
can, and M.E. Çalikan. 2006. Response of cotton (Gossypium hirsutum L.) to different tillage systems and intra-row spacing. Soil Tillage Res.
85
:221–228.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
Texas Agricultural Experiment Station Farm, Burleson County, Texas. Data combined over 2000 and 2002.
Texas Agricultural Experiment Station Farm, Burleson County, Texas.
