Published online 24 February 2006
Published in Crop Sci 46:886-892 (2006)
© 2006 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP BREEDING, GENETICS & CYTOLOGY
Effect of Genotype and Genotype x Nitrogen Rate Interactions on Color in Juice and Raw Sugar from Sugarcane
Phillip A. Jacksona,*,
Michael G. O'Sheab,
Allan R. Ratteyc,
Graham D. Bonnettd,
Patricia F. Lindemanb,
Mike C. Coxe,
Joan E. Vickersf and
Terry Morgang
a CSIRO Plant Industry, Davies Lab., PMB, PO Aitkenvale, Qld. 4814 Australia
b Bureau of Sugar Experiment Stations, Meiers Road Indooroopilly, Qld. 4068 Australia
c Bureau of Sugar Experiment Stations, Gordonvale, Qld. 4865 Australia
d CSIRO Plant industry, Queensland Bioscience Precinct, 306 Carmody Road, St Lucia, Qld. 4066 Australia
e Bureau of Sugar Experiment Stations, Bundaberg, Qld. 4670 Australia
f Dep. of Biological and Physical Sciences, Univ. of Southern Queensland, West St, Toowoomba, Qld. 4350 Australia
g CSR Ltd, Kalamia Mill Estate, Ayr, Qld. 4807 Australia
* Corresponding author (phillip.jackson{at}csiro.au)
 |
ABSTRACT
|
|---|
There is an increased emphasis being placed on raw sugar quality in international sugar markets. Color intensity of raw sugar is one quality parameter of importance, with high color levels increasing refining costs and hence lowering sugar value. The quality of raw sugar is compromised by colored solutes or precursors of colored compounds. We examined genetic effects on color levels in juice from sugarcane (Saccharum spp.) in three field experiments, two of which also included different N fertilizer rate treatments. The relative impact on amino N in juice, an important precursor of colored compounds, was also examined. A key finding was consistently large variation among varieties (approximately threefold differences) for color and color-to-impurity ratio levels. These effects were highly repeatable across different environments and N rates, with genotype x environment and genotype x N rate interaction variance being not significant or less than 15% of genotype main effects in all experiments. Relatively small genetic effects were detected for amino N levels and these were less repeatable across environments than for color. Thus, selection for low color levels will be effective in sugarcane breeding programs. A relative color index is proposed to compare genotypes for propensity to produce color in raw sugar, considering the effects of both color and amino N in juice.
Abbreviations: CCS, commercial cane sugar • PPO, polyphenol oxidase
|
INTRODUCTION
|
|---|
COLOR ARISES in raw sugar because of a proportion of the solutes in sugarcane juice being present either as colored compounds or as precursors of compounds that are colored (Paton, 1992). Increased intensity of color in raw sugar leads to higher cost of removal of colored compounds during refining and therefore a potentially lower market value of the raw sugar. Color is therefore a key quality measurement in sugar industries and is defined and measured according to international standards (Chen and Chou, 1993).
Because of mixing and recycling of juice and product streams within sugar mills, experimentally it is difficult to relate raw sugar color to measurements made on specific batches of cane entering mills. However, the effects of some parameters may be inferred from an understanding of the manufacturing process. First, raw sugar is made to specified sucrose concentration (purity) in sugar mills. Therefore, the ratio of color intensity to total solute impurities will provide one indicator of propensity of cane juice to produce color in raw sugar, and a better indicator than color intensity in juice per se. Second, it is known from past studies that additional colored compounds are produced during sugar milling and sugar storage. The major additional source of colorants arises from the Mallard reaction, which is strongly affected by concentration of amino N level in juice (Paton and McCowage, 1987). Therefore, both color-to-impurity ratio and amino N content in juice are key parameters expected to affect raw sugar color.
Variation in juice color intensity between some sugarcane varieties has been reported (Abernethy and Aitken, 1986). Rates of fertilizer N are also known to affect the concentration of the color precursor amino N in sugarcane juice (Keating and Catchpoole, 1995; Keating et al., 1999). However, most (if not all) sugarcane industries up to now have given little or no consideration to color effects in variety selection or release decisions or in setting recommended N fertilizer rates. The objective of the work reported here was to determine the level of genetic variation in color and concentration of the color precursor amino N in sugarcane and repeatability of genetic effects across different N fertilizer inputs. Genetic populations representative of both early (unselected) stages of selection and final phase evaluation in a commercial sugarcane breeding program were examined. This information, when combined with information about the economic value of changes in color, could be used to help decide if and when selection for low color or amino N levels should be conducted in sugarcane breeding programs. Using additional information about color formation during milling, and during storage of raw sugar, we propose an index for estimating relative propensity of different varieties or fertilizer treatments to produce color in raw sugar.
|
MATERIALS AND METHODS
|
|---|
Field Experiments
Three field experiments were conducted and are described below.
Experiment 1
Fifty-two genotypes were taken at random from populations generated from biparental crosses made by the Bureau of Sugar Experiment Stations (BSES) and CSR Limited. Each genotype was derived from a different cross. The same set of genotypes were planted in experiments at two sites in the Herbert River valley, Queensland, Australia (both 145°E, 18.5°S) on 24 June and 26 Sep. 1999, respectively. At each site, a randomized complete block experimental design was used, with two replicates. The unit plot consisted of 4 rows x 10 m, with 1-m gap between the ends of the plots. At each site the crop was managed according to local conventional practices, including application of 200 kg/ha N and 80 kg/ha K following planting. The experiments were harvested on 23 Aug. 2000 (Abergowrie) and 30 Aug. 2000 (BSES). At the same time, a 6-stalk sample was taken from the middle two rows of each plot, and juice extracted by the small mill method. In this experiment and others reported in this paper, the tops of the stalk above the growing point were removed, along with all dead leaves. Measurements on each juice sample were made as described for other experiments below. In addition, measurements of the activity of the enzyme polyphenol oxidase (PPO) were made on the same juice samples. This enzyme is believed to be a major contributor to color development in sugarcane (Vickers et al., 2005). PPO was measured as oxygen uptake (µm min–1 mL–1), with an oxygen electrode following the method described in Vickers et al. (2005). In analysis of data, a split-plot model was used (Steel and Torrie, 1981) with environments and clones considered as random effects for the purpose of estimating variance components. Data for this experiment, and others reported in this paper, were analyzed by the PROC GLM procedure in the SAS statistical software package (SAS Institute, NC, USA).
Experiment 2
This experiment was planted in a split plot design with four N rates (0, 80, 160, 240 kg/ha) applied as whole plot treatments in each of two blocks and one plot of each of 12 genotypes randomized within each whole plot. Each genotype plot consisted of four rows x 10 m, with an interrow spacing of 1.5 m. The experiment was planted on 8 Oct. 1998 on a farm in the Isis sugar mill region (146°E, 25.4°S). The crop was grown following conventional cultural practices for the region. The experiment was sampled and harvested on 6 Sep. 1999 (plant crop), 18 Sep. 2000 (first ratoon crop), and 1 Oct. 2001 (second ratoon crop). Samples for color determination were made in the first and second ratoon crops. Juice was extracted from 6-stalk samples taken from the middle two rows of each plot at harvest, by the small mill method. A split plot design in space and time (Steel and Torrie, 1981) was used in analysis of data to identify statistically significant sources of variation.
Experiment 3
Two separate subexperiments were planted on 30 April 1999 adjacent to each other in the same field with one being harvested in July each year and the other in September. Harvesting dates were 3 July 2000 and 11 Sep. 2000 in the plant crop, and 2 July 2001 and 10 Sep. 2001 in the first ratoon crop. Time of harvest effects are therefore confounded with possible environmental effects in the field used. Within each subexperiment (harvesting schedule), a split plot design was used, with three N rates (0, 125, 280 kg N/ha in the plant crop, and 0, 179, and 358 kg N/ha in the first ratoon crop) applied as whole plot treatments within each of three blocks. Fifteen genotypes were randomized within each of the whole plots. These 15 genotypes were either cultivars in the Burdekin region or advanced stage selections being considered for possible release in the region. The unit plot size was 4 rows x 10 m, with 1.5-m interrow spacing. In 2000, color related measurements were made only in the July harvested subexperiment, while in 2001, measurements were made in both the July and September harvested subexperiments. Juice was extracted by crushing fibrated cane using a hydraulic press (McRae et al., 1996), and measurements on each juice sample were made as described in the sections below. Analyses were performed on each of two separate subsets of these data: (i) results from the July samplings in both 2000 and 2001 and (ii) results from the July and September samplings in 2001. A split plot design in space and time (Steel and Torrie, 1981) was used in analysis of data to identify statistically significant sources of variation.
Measurement of Pol, Brix, and Impurities
Following juice extraction from stalks, a juice sample of about 0.5 to 1 L was taken and mixed before immediate measurement of brix and pol. Brix is a measure of total solutes in juice, while pol is an industry standard measure of sucrose in juice (BSES, 1984). A subsample of about 200 mL was then taken, frozen quickly with dry ice, and then kept frozen until color measurements were subsequently made.
Brix (expressed as g solute per 100 g of solution) was measured with a refractometer. Pol (g sucrose/100 g of juice) was estimated from:
This formula was derived from regression analysis of data in Table 1 in BSES (2001).
View this table:
[in this window]
[in a new window]
|
Table 1. Color (I.C.420) at pH 4, 7, and 9, color (at pH 7) at 1% impurity, and total amino nitrogen (TAN) for four N rates for Exp. 2. Twelve genotypes representative of advanced stages of selection in a sugarcane breeding program were evaluated across four N rates in first and second ratoon crops. Averages across all 12 genotypes are given.
|
|
The polarimeter readings were made on Schmidt and Haensch instruments calibrated so that a "normal" mass of 26.00 g of pure sucrose in 100 mL of solution will read 100 degrees (Chen and Chou, 1993).
Impurities per mass of juice (grams impurities/100 g juice) was determined from:
Purity was determined from:
Color Measurements
Following thawing, the brix of the sample was recorded again. A 5-mL aliquot was made up to 100 mL with MilliQ water. Approximately 1 g of Kieselghur was added and the flask shaken. The sample was then filtered and the filtrate divided into three separate beakers and pH adjusted to 4.0, 7.0, and 9.0 with HCl and NaOH. The absorbance was immediately read in a 1-cm tube at 420 nm. The concentration of brix in units of g/mL was estimated from BSES (2001; Table XII). A color index, following standard procedures (Chen and Chou, 1993) termed simply "color" in this paper was determined from the following:
This value has been reported as ICUMSA units (I.C.x; Chen and Chou, 1993) where x = wavelength used in measuring absorbance (in nm). Comparisons of impurity levels such as color and ash in juice entering a sugar mill are more appropriately presented on a per unit impurity basis when considering impact on sugar quality, because raw sugar is generally produced to a constant level of total impurities (e.g., Abernethy and Aitken, 1986). Color was divided by (100 – purity) and this ratio was termed "color at 1% impurity," to provide an index of color on a per unit impurity basis.
Total Amino–Nitrogen Analysis
Total amino–nitrogen analyses were determined with an automated system reported by Chapman et al. (1996). Aspartic acid solutions were run as external standards at concentrations of 5, 10, 15, 20, and 30 µg/mL. Final concentrations from unknown samples were determined from the calibration curve and calculated taking into account the dilution factors used. The results were expressed here in units of "mg aspartic acid/kg brix" for each sample. Conversion to total amino N/kg brix can be done by multiplying by 14 (molecular weight of N) and dividing by 133.1 (molecular weight of aspartic acid).
|
RESULTS
|
|---|
Experiment 1
There was highly significant and large variation among genotypes for color, color at 1% impurity, and PPO activity (Table 2). Among the experimental (non cultivar) genotypes, there was approximately a three-fold difference in these measurements between the lowest five and highest five genotypes and genetic coefficients of variation of 30% or greater, indicative of high genetic variance. Interestingly, most of the cultivars included in the trial had color levels near to or less than the lowest experimental genotypes. For all three measurements, genotype x environment interaction variance was low compared with genetic variance (less than 15%), indicating stability in ranking of genotypes across the two sites. This occurred despite contrasting responses in growth as shown by large genotype x site interactions for cane yield (Table 2). Collectively, these results suggest that large gains in improvement of color levels in unselected populations of genotypes in breeding programs could be attained via selection.
View this table:
[in this window]
[in a new window]
|
Table 2. Variance components for genotypes, genotype x site interaction, and error variance in Exp. 1 for color (I.C.420) at pH 7, color at 1% impurity, polyphenol oxidase activity (PPO), and cane yield (t/ha). This experiment comprised a sample of clones representative of the early stages of selection and standard cultivars as indicated, evaluated across two sites. Means of the bottom and top ranked five genotypes (averaged across the two sites), and performance of each standard cultivar (averaged across the two sites) are also shown for each trait.
|
|
Genetic correlations (± standard errors) between color and the traits: color at 1% impurity, PPO activity, cane yield, and CCS, and impurities were 1.02 ± 0.12, 0.62 ± 0.11, –0.42 ± 0.20, –0.17 ± 0.15, and 0.19 ± 0.12, respectively. The negative correlation between color and cane yield is of interest, without any obvious physiologically based explanation. However, the practical consequence of such a correlation is that selection for high cane yield would tend to indirectly reduce color levels in such populations. The high correlation between color and color at 1% impurity, without any correlation of note observed between color and impurities, is indicative of the predominant effects of genetic variation in color, rather than variation in impurity level, in driving variation in color–impurity levels. The strong correlation between color and PPO activity is consistent with the hypothesis that PPO has an important role in developing color in juice (Vickers et al., 2005) but does not prove such a causal relationship.
Experiment 2
There was a high correlation between color measured at different pH levels, across genotypes and N fertilizer levels (Fig. 1
, Table 1). A similarly high correlation between color measured at different pH levels across treatments was also observed in all other experiments (data not shown for other experiments). Thus, color measured at pH 7 provided a close guide to the relative effect of different genotype and N fertilizer treatments on color in juice overall. This result suggests either or both the following: (i) that common causal colorants predominate in causing color at different pH levels, or (ii) if different causal colorants are primarily involved in causing color at different pH levels, these are highly correlated across treatments.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 1. Color (I.C.420) at pH 4 and 9, and color (at pH 7) at 1% impurity, versus color at pH 7, across 12 genotypes in Exp. 2. This experiment evaluated genotypes across three contrasting N rates, and the average of genotypes across all N levels is shown. Vertical and horizontal bars indicate least significant differences (P < 0.05).
|
|
Increasing N rates were associated with significantly decreasing (P < 0.05) color and color–impurity levels (Table 1). Nitrogen rate had no significant effect on amino N levels and there was no N rate x crop-year interaction observed for any measurement.
A large range in both color and color–impurity ratio was observed for genotypes, with a greater than two-fold range between lowest and highest varieties in both crop years (Fig. 1). However, genotype x N rate and genotype x crop-year interactions were not significant for color or color–impurity ratio. Color varied much more among genotypes compared with impurity level, leading to a high correlation between color and color–impurity ratio (Fig. 1).
Experiment 3
Nitrogen rate and genotypes had highly significant (P < 0.01) effects on color, color at 1% impurity and amino N levels. Interaction effects were not significant, with the exception of genotype x year effects and N rate x year interaction for amino N and genotype x N rate interaction for amino N in 2001.
Increasing N fertilizer rate resulted in a large increase in amino N levels but a decrease in color and color–impurity levels (Table 3). The variation in color and color–impurity levels paralleled each other, indicating that the effect of N level on color was the primary driver of variation in color–impurity levels rather than impact on impurity levels.
View this table:
[in this window]
[in a new window]
|
Table 3. Color (I.C.420) at pH 7, color index at 1% impurity, and total amino N, for Exp. 3 at three times across plant (2000) and first ratoon (2001) crops and three N rates. Fifteen genotypes including commercial varieties and advanced stage selections were evaluated at three levels of N fertilizer and the values given represent the averages across all genotypes.
|
|
There was a large level of variation among genotypes for both color and color at 1% impurity (Fig. 2
). There was greater than a two-fold difference between the lowest color genotypes (Q124 and Q208A) and those with the highest color (Q96, Q189A and 94A6030). Color and color at 1% impurity were strongly correlated across genotypes (Fig. 2), indicating that color rather than impurity levels were the main factor responsible for variation among genotypes in color/impurity levels. Although genotype x year interactions were found to be significant in these data, this source of variation was small relative to the large main effect of genotypes, resulting in the high correlation in relative genotype ranking for color and color–impurity between years (Fig. 2).
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 2. Performance of genotypes in Exp. 3 for: (a) Color (I.C.420) in 2000 versus 2001, (b) color at 1% impurity in 2000 versus 2001, (c) color at 1% impurity versus mean color averaged across 2000 and 2001, and (d) amino N levels in 2000 versus 2001. In this experiment genotypes were tested across three N rate treatments and two time of harvesting treatments in a plant crop (2000) and ratoon crop (2001). Genotype x N rate or time of harvesting effects were not significant and the average of genotypes across treatments in each year are shown here. Vertical and horizontal bars indicate standard errors.
|
|
Variation in amino N levels was also large between genotypes (Fig. 2), with around a two-fold variation in levels between genotypes in both years. However, unlike color and color at 1% impurity, there was considerable reranking of genotypes between the years. This was possibly associated with the large differences in N status of the crop between years, indicated by contrasting average amino N levels between years (Table 3). This result suggests that some genotypes have a greater capacity to accumulate excess N as amino N than others under high N levels.
|
DISCUSSION
|
|---|
A key result in this research was the very large genetic variation revealed for color and color–impurity levels in sugarcane. This variation was apparent in genotypes representative of both early and final selection stages in breeding programs and was consistent across the three experiments. Some genotypes were identified with up to three-fold higher color–impurity ratios compared with some existing commercial cultivars. Color measurements across genotypes also demonstrated a high level of repeatability between different environments, including different sites, different crops and years, and different N rates. The results collectively illustrate a high potential for varieties to affect color intensity in juice entering sugar mills and also scope for large gains from selection for clones with juice of lower color intensity at all stages of selection. However, there was generally less genetic variation for amino N levels, suggesting this component of high color sugar would be less easily manipulated by genetic selection.
Increasing rates of applied N had little but generally negative impact on color in juice and color–impurity levels but increased amino N levels. Overall, there was a negative relationship between color-to-impurity ratio and total amino N level for different N rates. Thus, increased N fertilizer rate in this experiment may be expected to have two different effects on raw sugar color, each effect acting in the opposite way. The expected net impact of these two effects on raw sugar color is discussed below. The increase in amino N levels with increasing N rate was expected and consistent with previous research (e.g., Keating et al., 1999), but the effect on color and on color–impurity has not been reported before to our knowledge.
It is important that results obtained relating to impact of varieties and fertilizer rates on color and amino N are interpreted in terms of likely economic impact before application in breeding programs or farm management. In estimating these economic effects, it is important to consider two issues: (i) the net effect on raw sugar color–impurity and (ii) the effect of changes in raw sugar color–impurity on its market value. The latter is not discussed in this paper because this is industry specific and commercially sensitive, but possible approaches to the first issue are discussed below.
Color in raw sugar when delivered to refiners results from numerous processes, and it is not possible to precisely predict the intensity of color in raw sugar from juice parameters and milling process specifications. This complicates prediction of how varieties or N rates will affect color intensity of raw sugar and hence industry profitability. However, use of best available knowledge to guide decisions should usually lead to a more economically optimal outcome, compared with adopting arbitrary or extreme positions on the basis of no analysis. Extreme positions in relation to color management via varieties or N rate recommendations could include ignoring color in selection or recommendations, consistent with variation in color having very little or no expected economic impact. By contrast, an extremely cautious position could also be adopted, for example, restricting selection and release of all varieties with significantly higher color than existing cultivars.
Two different methods of juice extraction were used in different experiments during this research, namely the small mill method and the press method (McRae et al., 1996). Prior investigations (not reported here) showed that different extraction methods gave the same information about relative effects of genotypes and N, so that results and conclusions were not sensitive to method of extraction.
The following basic assumptions are used here to derive an index to predict how juice with different color intensity and levels of amino N affects color intensity of raw sugar.
Raw sugar is made to a specified narrow purity range, and, therefore, the ratio of color–impurities in massecuite and raw sugar determines the total amount of colorant in the raw sugar.
Color arises from (i) colorants present in cane before crushing, (ii) colorants produced by enzymatic or non enzymatic oxidation soon after crushing, (iii) colorant produced from the Mallard reaction between reducing sugars and amino N, (iv) colorant from alkaline degradation products of reducing sugars, and (v) colorant produced from caramelization reactions (Paton and McCowage, 1987; Paton, 1992). Of these (i), (ii), and (iii) are the most important sources of colorant (Paton, 1992).
The Mallard reaction occurs only at high brix levels and proceeds at a rate which is approximately linearly related to amino N levels (Paton and McCowage, 1987).
There is a high correlation between color intensity measured in first expressed juice and final mixed juice in sugar mills (Abernethy and Aitken, 1986).
Given that the major sources of colorant in raw sugar are (i) those present soon after crushing and (ii) those arising from the Mallard reaction, some function of color, amino N, and impurities in juice should provide a reasonable estimate of relative propensity of different juices to produce color in raw sugar. However, a key factor is the relative importance of these two sources, to derive an index that appropriately weights measurements of color in first expressed juice, and amino N levels.
Published reports (Mundey et al., 1968), and our own observations (data not shown), suggest that color developed during milling represents between approximately 40 and 100% of color present in clarified juice. It is also known that the Mallard reaction increases color in sugar stored before delivery to the customer. Cortis-Jones et al. (1973) showed that the color stability of raw sugar was closely and linearly related to the amino N concentration, with the mean rate of increase in color intensity across a wide range of sugar samples with different amino N levels directly proportional to amino N concentration. The average relative increase in color intensity in sugar storage in the Australian sugar industry is difficult to determine, but for the purposes of this paper we have assumed an average increase of 15% on the basis of input from sugar industry personnel with expertise related to sugar storage.
On the basis of the above, it is assumed that an increase in color of 60% occurs during milling and storage. From this, the following relative color index (RCI) is proposed to estimate relative propensity of different treatments (e.g., varieties or N rates) to produce color in raw sugar:
 |
where relative color intensity, amino N, and impurities are measured in first expressed juice and expressed as a percentage of the average of treatments (e.g., varieties, N rates) being compared.
The above index can be used toward predicting economic impact due to color changes associated with different varieties. Results from commercial cultivars included in Exp. 3 are used below to illustrate the level of variation. Table 4 shows the major cultivars in the Burdekin sugarcane growing region at present. A RCI is determined on the basis of the formula above. If it is assumed that the average RCI of all varieties in the Burdekin is currently 1400, then the impact of adoption of an extremely high color variety such as Q189A which has a RCI of 50% greater than this average can be determined. The economic impact of this change may be compared with potential benefits which the variety may bring in relation to other traits (e.g., improved cane yields), and a similar approach could be taken for developing optimal N rates.
View this table:
[in this window]
[in a new window]
|
Table 4. Color index at 1% impurity, amino N levels and relative color index (RCI) for a subset of genotypes in Exp. 3. Values are based on the average across all N rates and illustrate the range of variation in RCI in a set of commercial cultivars.
|
|
|
ACKNOWLEDGMENTS
|
|---|
This work was funded in part by the Sugar Research and Development Corporation, Australia. We acknowledge with thanks the competent assistance of technical staff from CSIRO, BSES, and CSR Ltd in conducting the field trials and collecting samples for this research, with particular thanks to Bill Messer and John Foreman. We thank CSR Ltd for allowing us to take samples in their sugar mills.
|
NOTES
|
|---|
This research was part funded by the Sugar Research and Development Corporation.
Received for publication July 21, 2005.
|
REFERENCES
|
|---|
- Abernethy, P.E., and A.T. Aitken. 1986. Factors affecting the levels of color entering a sugar mill. Proc. Aust. Soc. Sugar Cane Technol. 8:1–7.
- BSES. 1984. Laboratory manual for Australian sugar mills. Volume 1. Principles and practices. Bureau of Sugar Experiment Stations, Indooroopilly, Qld. Australia.
- BSES. 2001. Laboratory manual for Australian sugar mills. Volume 2. Analytical methods and tables. Bureau of Sugar Experiment Stations, Indooroopilly, Qld. Australia.
- Chapman, L.S., G. Ham, A.P. Hurney, D.H. Sanders, and G.J. Leonard. 1996. Preliminary evaluation of field factors affecting amino acids in sugarcane juice. Proc. Aust. Sugar Cane Technol. 18:212–221.
- Chen, J.C.P., and C.C. Chou. 1993. Cane sugar handbook. 12th ed. John Wiley & Sons. New York.
- Cortis-Jones, B., E. Forsyth, and H. Hamilton. 1973. Field factors and color of raw sugars at shipment. CSR Research Laboratories, Sydney.
- Keating, B.A., and V.R. Catchpoole. 1995. Sugarcane yield and nitrogen uptake in relation to profiles of mineral-nitrogen in the soil. Proc. Aust. Soc. Sugar Cane Technol. 17:187–192.
- Keating, B.A., G. Kingston, A.W. Wood, N. Berding, N, and R.C. Muchow. 1999. Monitoring nitrogen at the mill to guide N fertilization practice on farm. Proc. Aust. Soc. Sugar Cane Technol. 21:10–19.
- McRae, T.A., J.K. Bull, B.G. Robotham, and R.C. Sweetham. 1996. Measuring sugar content in variety trials. In J.R. Wilson et al. (ed.) Sugarcane: Research towards efficient and sustainable production. CSIRO Division of Tropical Crops and Pastures. Brisbane, Australia.
- Munday, B.M., T.G. Burgess, R.V. Ames, and C.W. Davis, C.W. 1968. Color in raw sugar manufacture. Proc. Int. Soc. Sugar Cane Technol. 13:395–404.
- Paton, N.H. 1992. The origin of color in raw sugar. Proc. Aust. Sugar Cane Technol. 14:8–17.
- Paton, N.H., and R.J. McCowage. 1987. Color forming mechanisms during mill processing. Proc. Aust. Soc. Sugar Cane Technol. 9:11–20.
- Steel, R.G.D., and J.H. Torrie. 1981. Principles and procedures of statistics. McGraw-Hill International Book Company, London.
- Vickers, J.E., C.P.L. Grof, G.D. Bonnett, P.A. Jackson, D.P. Knight, S.E. Roberts, and S.P. Robinson. 2005. Over-expression of polyphenol oxidase in transgenic sugarcane results in darker juice and raw sugar. Crop Sci. 45:354–362.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
