Phenotypic Evaluation of Saccharum spp. Genotypes during the Plant-Cane Crop for Biomass Production in Northcentral Mississippi ^†

Abstract

Saccharum is relatively new to 33° N latitude. S. spontaneum readily hybridizes with commercial sugarcane (Saccharum spp.) and lends cold tolerance and greater yield to the hybrid progeny, called energycane. Since 2007, there have been numerous new hybrid and backcross energycane genotypes developed but there is a paucity of information about them. Twenty energycane genotypes were tested in the first season of growth from cane propagules (plant cane; PC) against Ho 02-113 (a control) for two site-years in northcentral Mississippi. Grand (exponential) growth continued into October. The prevailing paradigm is that tonnage is what matters. Except for percentage cellulose, all factors tested (dry matter yield, extractable juice volume, °Brix, theoretical ethanol from fermentation, theoretical ethanol from cellulose, and total theoretical ethanol) were greater from the second site-location compared to the first. Dry matter yield (DMY) and total theoretical ethanol yield (TTEY) were moderately correlated. Over the two years of this test only Ho 14-9213 exceeded in mean DMY of Ho 02-113. Sixteen of the 19 test genotypes in this test equaled or exceeded the mean TTEY of Ho 02-113.

1. Introduction

Perennial grasses can produce renewable biomass with fewer inputs than annual crops due to their ability to re-emerge for multiple seasons from a single planting. Sugarcane (Saccharum spp.) has potential as a bioenergy crop, but its tropical origin and lack of cold tolerance limits production from cooler latitudes. The S. spontaneum genotype US 56-15-8 originally from Thailand, as part of the USDA-ARS Sugarcane Research Unit’s basic sugarcane breeding program was imported to provide resistance to sorghum mosaic virus (SrMV; family Potyviridae) to commercial sugarcane. Hybrids of sugarcane × S. spontaneum were backcrossed to sugarcane to obtain SrMV resistant sugarcane lines [1]. Saccharum backcross progeny segregated into low-fiber, high-sugar lines, and higher-fiber, low-sugar, cold hardy lines called energycane. The advantage of Saccharum as a biomass crop in the United States is attributed to yields that exceed other biomass crops. Energycane yields in Louisiana exceed 100 Mg ha⁻¹ fresh cane weight. A study found the composition of bagasse from energycane to be 43% cellulose, 24% hemicellulose, and 22% lignin [2]. The breakdown of cellulose and hemicellulose into fermentable monomers of glucose and xylose can be used in the production of second-generation biofuels [3]. Studies [2,4] estimated ethanol production from sugarcane to be between 102.9 to 128.7 g L⁻¹ juice. Energycane’s higher production of fiber led to higher estimated ethanol production than sugarcane. Greater net energy ratios of energycane give it an advantage to even elite sugarcane varieties [5,6,7]. Addition of a new crop to agriculture enhances crop diversity. However, testing is necessary in the areas north of Louisiana as they exceed the latitudes (30° N) where cane is traditionally produced.

The genus Saccharum is genetically complex. Owens [8] suggests that because of a relatively high lignocellulosic content and total tonnage, energycane is a prime candidate for bioenergy. Germplasm derived from crosses between commercial sugarcane and wild S. spontaneum are usually associated with a lower sucrose yield and greater fiber content than commercial sugarcane varieties [9,10,11]. Such crosses capitalize on interspecific hybrid vigor which can be maintained when planting vegetative cuttings [12,13]. Because of wide hybridization and hybrid vigor, energycane is associated with a high biomass yield per unit land, increased disease resistance, and cold weather tolerance [14,15]. This affords energycane a greater potential climate range than current commercial sugarcane cultivars [13]. Preliminary data were collected on energycane from 2007 to 2013 as one of the five herbaceous species in the Herbaceous Feedstocks Partnership [16]; Owens [8]. Five energycane genotypes, obtained from the USDA-ARS Sugarcane Research Unit at Houma, LA, were tested for yield stability over six cropping years at eight locations in the Continental U.S. and Hawaii: the northern most location was Starkville, MS (33.4504° N; [17]. Since 2007 researchers at the ARS Sugarcane Research Unit have continued to develop new energycane genotypes, but there is little to no information about these new genotypes. Energycane’s genetic diversity, tonnage potential, and ability to ratoon with consistent yields for up to five years with minimal nutrient requirements make it a candidate as a biomass crop for bioenergy production [11,13]. In this study we asked the questions: Do new genotypes meet or exceed the same survival and yield as the control (Ho 02-113)? Do they vary meaningfully in other traits used for bioenergy? To answer these questions, we tested 20 new genotypes at the same extreme north location used in 2018.

The objective of this study was to determine if any of these 20 genotypes exceed the abilities of Ho 02-113 (control) at 33° N latitude to produce fermented and cellulosic ethanol. To assess total potential ethanol production, yield, °Brix, and percentage cellulose of these 20 new energycane genotypes were measured.

2. Materials and Methods

2.1. Planting Material and Study Site

The germplasm used in this study was provided by the USDA-ARS Sugarcane Research Unit in Houma, LA. It was originally selected for cold tolerance in a multi-year, unreplicated study under freezing conditions in Louisiana. Genotypes were assigned designations beginning with the international call sign “Ho” to designate their origin in the USDA-ARS breeding program in Houma, LA, and are listed in

Table 1. The 28 genotypes planted at the R.R. Foil Plant Science Research Center (Starkville, MS) during the fall of 2018 to assess survivability, yield, and theoretical ethanol production.

Original 28 Genotypes
Ho 15-9901	Ho 14-9218	Ho 15-9915	Ho 15-9922
Ho 15-9902 *	Ho 14-9213	Ho 14-9228 *	Ho 15-9923
Ho 15-9903	Ho 14-9220 *	Ho 15-9917 *	Ho 15-9924
Ho 15-9904	Ho 15-9911	Ho 15-9918	Ho 14-9276
Ho 15-9905	Ho 15-9912	Ho 15-9919	Ho 15-9926 *
Ho 15-9906	Ho 15-9913	Ho 15-9920	Ho 02-113
Ho 15-9907	Ho 15-9914 *	Ho 15-9921	L 01-299 *

* Asterisk indicates genotype that was not carried on to the replicated trial at H.H. Leveck Animal Research Center (2020) nor the Bearden Dairy Research Center (2021).

. In September 2018, canes of 26 genotypes, the energycane control Ho 02-113, and the sugarcane control L 01-299 (Table 1) were planted in an increase nursery at R.R. Foil Plant Science Research Center (near Starkville, MS; 33.467952, −88.754920). This location is 345 km north of typical sugarcane production areas. The genotypes were planted as canes in rows 6.1 m long and 1.8 m on center as whole stalks topped below the apical meristem. Canes were laid in every other furrow overlapping by approximately one third of each cane length The soil at this location was a Leeper silty clay loam classified as fine, smectitic, nonacid, thermic Vertic Epiaquepts. All subsequent seed cane for this study came from these plots.

Twenty of these genotypes and the energycane control (Ho 02-113) were selected for further testing in the fall of 2019. The other six genotypes and the sugarcane control (L01-299) were not carried forward for further testing due to lodging, low yield, or failure to survive at Starkville, MS. The selected genotypes and energycane control were planted (3 October 2019) in a randomized complete block design with four replications at H.H. Leveck Animal Research Center in the fall of 2019 (33.423582, −88.792412;

Table 2. Field locations, planting, and harvest years of tests conducted during the plant cane year of each field.

Site	GPS Coordinates	Cane Planting Date	Origin of Material	Plant Cane Harvest Year	Field Plot Design
R.R. Foil Plant Science Research Center	33.457952 −88.754920	3 September 2018	USDA Sugarcane Research Center	3 October 2019, 25 September 2020	Unreplicated nursery
H.H. Leveck Animal Research Center	33.42582 −88.792415	4 October 2019	R.R. Foil Plant Science Research Center	25 September 2020	Randomized complete block
Bearden Dairy Research Center	33.39499 −88.74101	27 September 2020	R.R. Foil Plant Science Research Center	25 November 2021	Randomized complete block

). The soil at this site-year is a Catalpa silty clay loam, somewhat poorly drained, classified by fine, smectitic, thermic Fluvaquentic Hapludolls. Beds and furrows were spaced 0.92 m apart in the fall prior to the first harvest. Seed cane was hand-harvested from the increase nursery located at the R.R. Foil Plant Science Research Center. Beds, and therefore furrows, were spaced 0.92 m apart. Seed canes were harvested using a machete. The distal end of the cane was removed so that each planted cane measured either 1.83 m, 2.44 m, or 3.05 m depending on height of the canes. To obtain a uniform node count, canes that were 1.83 m long, were planted 5 to a 6.1 m plot; those that were 2.44, seven to a plot; and those that were 3.05 m long planted six to a plot. Stalks were laid in every other furrow overlapping by approximately one third of each cane length resulting in plots that were 1.8 m apart on center. They were subsequently covered with soil using a three-point tilt scraper blade. During the spring after planting 48 20 cm deep soil cores were taken and analyzed for soil nutrients. Nitrogen (UAN 30-0-0) was applied at a rate of 168.1 kg N ha⁻¹ using a knife rig on 17 June 2020.

The same 21 genotypes were carried into a new replicated field test established in the fall of 2020, located at Bearden Dairy Research Center (33.39499, −88.74101, and 82 m elevation) near Sessums, MS using the same harvest and planting techniques employed in the 2019 planting. Soil at Bearden Dairy was a Kipling silty clay loam classified as fine, smectitic, thermic Vertic Paleudalfs.

2.2. Data Collected

2.2.1. Stalk Heights and Counts

Stalk heights and stand densities were recorded biweekly starting on 12 June 2020 and ending on 16 October 2020 at the H.H. Leveck Animal Research Center. Mean heights were plotted to visualize relative growth and rate of growth, and to determine the onset and cessation of the grand growth period. Heights were determined based on the mean height of five random stalks per plot. Heights were measured (cm) from the ground to the most distal set of auricles. Stand densities were obtained by counting each stalk found in each plot and extrapolating to stalks ha⁻¹. Similar data were recorded in 2021 at the Bearden Dairy Research Center; however, a slightly earlier onset of growth shifted the stalk height and stem density counts to start on 10 May and terminate on 12 October 2021.

2.2.2. Biomass Composition and Yield

Plots were harvested and weighed in 2020 and 2021 at the end of the season using a Cibus S Wintersteiger plot harvester (Ried im Innkreis, Austria) to determine fresh weight yields (FWY) (Mg ha⁻¹). Weight of all destructive harvests prior to end-of-season harvest were added back to the plot harvest weight. A sub-sample of chopped stalks (1 kg or greater) from each genotype was collected, weighed in the field, and dried for moisture determination according to Equation (1). The dried field subsamples were subsampled again. These were ground to pass through a 2 mm screen and submitted for analysis. These samples contained residual soluble carbohydrates and therefore represent an accurate representation of cellulose as a percentage of the total sample weight. Percentage cellulose determination was made by the H.W. Essig Nutrition Lab using van Soest detergent fiber analysis 1963 [18] and was calculated by subtracting the soluble carbohydrate and lignin values in the ground sample from the acid detergent fiber value. Data from the 2020 PC year were compared to the 2021 PC year.

Although high-fiber content corresponds to low sugar, and visa versa, the greater fiber percentages found in energycane affect juice volume (JV); which is an important consideration when considering ethanol from sugar fermentation. There is a tradeoff of ethanol from sugar fermentation and ethanol derived from cellulose conversion. Both these characteristics impact total ethanol production. Saccharum-based ethanol is produced by yeast-fermentation of the dissolved carbohydrates (measured as part of °Brix) contained within the juice.

Equation (1):

$$\mathrm{Percent} \mathrm{Moisture}={\displaystyle \frac{\left(\mathrm{Wet} \mathrm{Stalk} \mathrm{Weight} \left(\mathrm{g}\right)-\mathrm{dry} \mathrm{stalk} \mathrm{weight} \left(\mathrm{g}\right)\right)}{\mathrm{Wet} \mathrm{Stalk} \mathrm{Weight}}}\times 100$$

(1)

Equation (2):

$$\mathrm{Dry} \mathrm{Matter} \mathrm{Yield}={\displaystyle \frac{\mathrm{Fresh} \mathrm{Weight} \left(\mathrm{Mg}\right)}{\mathrm{ha}}}\times \left(100-\%\mathrm{moisture}\right)$$

(2)

To obtain °Brix (and extractable juice volume) five randomly sampled, “millable” stalks were removed from each plot to determine stem morphology—height (cm), weight (kg), and diameter of the lowest internode (mm). A “millable” stalk was defined as a stalk that had senesced leaves (not a recent tiller), and whose height was mean for the plot (dewlap reached the canopy). These sample stalks were crushed in a three-roller electric sugarcane juicer (Plant Based Pros^®; Jersey City, NJ, USA). For each stalk, the total juice volume was measured (mL) and subsequently stirred before determining °Brix with a digital refractometer. After juicing, crushed stalks were weighed, dried to a constant weight, and weighed again. Total extractable soluble solids were based on the percentage moisture multiplied by °Brix. This value was subtracted from dry matter yield (DMY) to obtain a “corrected” DMY for theoretical ethanol from cellulose (TEC) calculations. The use of two separate samples, and the subtraction of soluble solids from the total DMY was calculated to avoid an exaggerated estimate of fiber on a percent cane basis due to residual soluble solids following the extraction of juice.

Equation (3):

Total Soluble Solids = °Brix × %Moisture × DMY

(3)

Equation (4):

Corrected DMY = DMY − Total Extractable Solids

(4)

2.2.3. Theoretical Ethanol Yield

To predict theoretical ethanol yield from juice fermentation (TEF) of carbohydrates in extracted juice using this method, Equation (5) was used. °Brix is a measure of soluble solids such as sugar, salt, minerals, and proteins in cane juice. In sugarcane, this is predominately a measure of soluble sugar. As such, a simple multiplication of juice measured by Equation (3) °Brix can be used to estimate the production of total yeast-fermentable ethanol from a field of a particular genotype.

The amount of juice (kL ha⁻¹) was calculated using the moisture percentage obtained from Equation (3) and multiplied by the mean °Brix per plot to obtain soluble solids ha⁻¹. To account for non-fermentable solids, it was assumed 75% of °Brix was fermentable sugars (Wortmann et al., 2010 [19], therefore the °Brix was multiplied by 0.75 (Equation (4)). The product was then multiplied by 0.581 according to the stoichiometry of yeast fermentation. The equation for first-generation TEF was the following:

Equation (5): Theoretical Ethanol Fermentation (TEF) [19]

$${\displaystyle \frac{\mathrm{Theoretical} \mathrm{Ethanol} \left(\mathrm{L}\right)}{\mathrm{ha}}}={\displaystyle \frac{\mathrm{Juice} \left(\mathrm{kg}\right)}{\mathrm{ha}}}\times \left(°\mathrm{Brix} \times 0.75\right)\times 0.581$$

(5)

To predict total ethanol from cellulose (TEC) the corrected DMY (Mg ha⁻¹) was multiplied by percentage cellulose then by 174.2 to obtain ethanol L ha⁻¹ (Equation (6)). The constant, 174.2 is based on current hydrolytic cellulosic ethanol technology Dias et al., 2012 [8]. The equation for TEC calculation was as follows:

Equation (6) [20]:

Corrected DMY (Mg ha⁻¹) × %cellulose × 174.2 = L ha⁻¹

(6)

Total theoretical ethanol yield (TTEY) simply requires the addition of the TEF from juice and the TEC from cellulose in dry matter [20]. Mean percentage cellulose was determined using a subsample (1 kg or greater) from the plot harvest, dried, ground to homogeneity, and subsampled again. The subsamples were submitted to the H.W. Essig Nutrition Lab for van Soest detergent fiber analysis [18]. The corrected DMY was used to avoid including soluble solids that are in the dry weight measurements of the field plot samples.

Data were analyzed using PROC GLM procedures for means separation at α = 0.05. Two years of data were assessed: the PC year at H.H. Leveck Animal Research Unit (harvested in 2020) and the PC year at the Bearden Dairy Research Unit (harvested in 2021). Analysis of variance indicated significant differences due to site-year (α = 0.05) for all characteristics measured, except percentage cellulose. Therefore, the data of each site-year were analyzed separately, except for percentage cellulose. Results of each site-year are presented separately.

Pearson correlation coefficients (r) and coefficients of determination (r²) were calculated between percentage cellulose and °Brix in juice; percentage cellulose and JV; percentage cellulose and TTEY; TEF and TTEY; TEC and TTEY, DMY and TEF; and cellulose yield and TTEY. Comparisons were assessed to determine any useful predictive associations.

Because of extensive variance around the mean of each genotype, characteristics (DMY, JV, °Brix, TEF, % cellulose, TEC, and TTEY) of each genotype were compared to the control (Ho 02-113) [21]. Ultimately, the purpose of a breeding program and its subsequent testing is to produce new germplasm that performs better (produces more ethanol in this case) than the existing varieties; thus, we compared to the control variety (Ho 02-113).

3. Results

Testing of these 20 genotypes was to determine yield and productivity among these new energycane genotypes against the control (Ho 02-113) for their ability to yield both fermentable and cellulosic ethanol. All 20 genotypes survived the winters at both site-locations. In general, growing conditions were better in 2021 than 2020. The mean of every characteristic measured; height, stalk population, DMY, JV, °Brix, TEF, TEC, TTEY (

Table 3. Mean values for dry matter yield (DMY), extracted juice volume (JV), °Brix, theoretical ethanol from fermentation of juice (TEF), theoretical ethanol from cellulose (TEC), and total mean theoretical ethanol yield (TTEY) for the plant cane year at the H. H. Leveck Animal Science Research Center in 2020.

Genotype Houma (Ho)	Mean Total DMY (Mg ha−1)	Mean JV (kL ha−1) *	Mean °Brix Value	Mean TEF (kL ha−1) **	Cellulose Percentage †	Mean TEC (kL ha−1) ‡	Mean TTEY (kL ha−1)
14-9213	3.97 f §	5.4 hi	11.8 a–f	0.247 f	0.36 d–f	0.811 a–c	1.058 c–f
14-9218	13.06 ab	16.9 a	10.8 c–g	0.791 a	0.40 ab	0.662 a–f	1.453 a–c
14-9276	15.17 a	10.9 b–g	13.5 ab	0.618 a–d	0.34 ef	0.652 a–f	1.066 c–f
15-9901	7.81 d–f	8.09 e–i	13.7 a	0.471 b–f	0.34 f	0.408 f	0.880 ef
15-9903	13.24 ab	11.7 b–f	12.8 a–c	0.674 a–d	0.37 a–f	0.614 b–f	1.288 a–d
15-9904	7.05 ef	4.8 i	10.0 d–g	0.211 f	0.37 b–f	0.777 a–d	0.987 d–f
15-9905	12.32 a–c	9.8 d–i	11.1 b–g	0.461 c–f	0.38 a–d	0.857 ab	1.318 a–d
15-9006	10.34 b–e	12.0 a–f	12.3 a–e	0.654 a–d	0.36 d–f	0.440 ef	1.094 b–f
15-9007	11.04 a–e	10.0 c–h	12.6 a–d	0.553 a–e	0.39 a–d	0.621 a–f	1.174 a–e
15-9911	10.76 a–e	8.8 e–i	9.0 g	0.323 ef	0.40 a–c	0.835 ab	1.158 a–e
15-9912	12.72 a–c	7.0 f–i	12.6 a–c	0.397 d–f	0.38 a–d	0.660 a–f	1.056 c–f
15-9913	14.02 ab	10.0 c–h	12.7 a–c	0.548 a–e	0.36 ef	0.651 a–f	1.199 a–e
15-9915	7.54 d–f	5.0 hi	11.2 a–g	0.250 f	0.38 a–d	0.506 c–f	0.756 f
15-9918	10.54 b–e	12.5 a–e	9.8 e–g	0.581 a–e	0.38 a–d	0.926 a	1.507 a
15-9919	10.10 b–e	11.1 b–g	12.2 a–f	0.590 a–e	0.39 a–d	0.808 a–c	1.398 a–d
15-9920	12.73 a–c	14.8 a–c	9.7 fg	0.649 a–d	0.41 a	0.473 d–f	1.122 b–f
15-9921	12.92 ab	15.8 ab	11.0 b–g	0.748 ab	0.38 a–e	0.757 a–d	1.505 ab
15-9922	11.51 a–e	14.5 a–d	11.7 a–f	0.737 a–c	0.36 d–f	0.613 b–f	1.350 a–d
15-9923	11.98 a–d	12.5 a–e	12.4 a–e	0.665 a–d	0.40 a–c	0.738 a–e	1.402 a–c
15-9924	8.36 c–f	6.04 g–i	12.2 a–f	0.332 ef	0.36 c–f	0.722 a–e	1.053 c–f
02-113(ck)	12.34 a–c	4.9 i	11.1 a–g	0.236 f	0.39 a–d	0.779 a–d	1.015 d–f
Mean	11.1	10.2	11.6	0.518	0.38	0.679	1.190
LSDα=0.05	4.47	5.1	2.6	0.278	0.36	0.308	0.376
p value	<0.0001	<0.0001	0.0217	<0.0001	0.0317	0.0435	0.0010

* (mean juice volume cane⁻¹) × (millable canes ha⁻¹).** Juice volume (L ha⁻¹) × (°Brix × 0.75) × 0.581 = theoretical ethanol (L ha⁻¹) [19]. ^† Mg ha⁻¹ × 174.2 = L ha⁻¹ [20]. ^‡ Mean percentage cellulose was not significantly different by year (p = 0.271). ^§ Means followed by the same letter are not significantly different at α = 0.05 level.

and

Table 4. Mean values for dry matter yield (DMY), extracted juice volume (JV), °Brix, theoretical ethanol from fermentation of juice (TEF), theoretical ethanol from cellulose (TEC), and total mean theoretical ethanol yield (TTEY) for the plant cane year at the Bearden Dairy Research Center in 2021.

Genotype Houma (Ho)	Mean Total DMY (Mg ha−1)	Mean JV (kL ha−1) *	Mean °Brix Value	Mean TEF (kL ha−1) **	Cellulose Percentage †	Mean TEC (kL ha−1) ‡	Mean TTEY (kL ha−1)
14-9213	15.95 ab §	45.1 a	12.7 c–g	2.493 a	0.36 d–f	0.914 a–c	3.407 a
14-9218	14.49 a–c	33.7 b	11.7 f–j	1.682 bc	0.40 ab	0.871 a–c	2.553 b–d
14-9276	15.20 ab	25.3 b–g	12.4 d–h	1.354 b–g	0.34 ef	0.840 a–d	2.194 b–g
15-9901	13.74 a–d	18.7 d–h	15.6 a	1.291 c–g	0.34 f	0.761 a–e	2.052 b–g
15-9903	13.45 a–d	22.0 c–h	12.4 d–h	1.178 c–g	0.37 a–f	0.813 a–d	1.991 c–g
15-9904	12.04 b–e	16.3 f–h	11.3 g–j	0.806 g	0.37 b–f	0.676 c–e	1.482 g
15-9905	10.94 c–e	19.2 d–h	13.7 a–e	1.170 c–g	0.38 a–d	0.675 c–e	1.845 d–g
15-9906	14.02 a–c	29.5 b–d	14.0 a–d	1.761 b	0.36 d–f	0.747 b–e	2.507 b–d
15-9907	16.95 a	31.9 bc	10.3 ij	1.430 b–f	0.39 a–d	1.007 a	2.437 b–d
15-9911	12.09 b–e	29.3 b–e	12.0 e–j	1.563 b–d	0.40 a–c	0.740 b–e	2.303 b–e
15-9912	16.34 a	18.2 e–h	10.5 h–j	1.076 fg	0.38 a–d	0.995 ab	2.071 b–g
15-9913	13.40 a–d	14.2 h	15.0 ab	0.906 e–g	0.36 ef	0.766 a–e	1.618 e–g
15-9915	9.65 de	16.8 f–h	14.5 a–c	1.033 d–g	0.38 a–d	0.584 de	1.617 e–g
15-9918	9.17 e	19.7 d–h	10.2 j	0.889 e–g	0.38 a–d	0.573 de	1.462 g
15-9919	13.92 a–c	27.3 b–f	13.2 b–g	1.572 b–d	0.39 a–d	0.850 a–c	2.422 b–d
15-9920	15.94 ab	33.6 b	12.1 e–j	1.775 b	0.41 a	0.957 ab	2.732 a–c
15-9921	15.05 a–c	36.4 ab	12.1 e–j	1.916 ab	0.38 a–e	0.861 a–c	2.777 ab
15-9922	14.55 a–c	26.6 b–g	13.2 b–f	1.502 b–e	0.36 d–f	0.770 a–e	2.272 b–f
15-9923	8.51 e	15.8 g–h	14.4 a–c	0.997 d–g	0.40 a–c	0.538 e	1.535 fg
15-9924	15.13 a–c	20.7 d–h	12.0 e–j	1.071 c–g	0.36 c–f	0.863 a–c	1.934 d–g
02-113(ck)	16.06 ab	16.1 e–h	12.2 d–i	0.874 fg	0.39 a–d	0.937 a–c	1.811 d–g
Mean	13.9	24.7	12.6	1.354	0.38	0.801	2.158
LSDα=0.05	4.2	11.1	1.9	0.619	0.36	0.308	0.747
p value	0.0010	<0.0001	<0.0001	<0.0001	0.0317	0.0047	<0.0001

* (mean juice volume cane⁻¹) × (millable canes ha⁻¹). ** Juice volume (L ha⁻¹) × (°Brix × 0.75) × 0.581 = theoretical ethanol (L ha⁻¹) [19]. ^† Mg ha⁻¹ × 174.2 = L ha⁻¹ [20]. ^‡ Mean percentage cellulose was not significantly different by year (p = 0.271). ^§ Means followed by the same letter are not significantly different at α = 0.05 level.

) was greater in 2021 than in 2020, except percentage cellulose, which was unchanged. These differences are highlighted in Figure 1, which reflect the mean height of the cane in both PC crops. Plants in 2021 were taller from ordinal day 260 through to the end of the season.

3.1. Height and Growth Curves

Mean plant height for 2021 was greater than 2020, starting at week 7 (17 September) and continuing to the end of season.

Grand growth was defined as the period during which the rate of growth is increasing. Based on the slope of the lines in Figure 1, it appears the mean growth of the genotypes was still in grand growth (though the slope is starting to level off) at the end of the growing season in 2020 and 2021.

3.2. Dry Matter Yield

Significant differences in total DMY were observed among genotypes for the PC year in 2020 (p < 0.0001; Table 3). The mean DMY across all genotypes was 11.1 Mg ha⁻¹. The greatest mean DMY was 15.175 Mg ha⁻¹ (Ho 15-9925), and the least was 3.975 Mg ha⁻¹ (Ho 15-9909). The DMY of the control, Ho 02-113, was 12.342 Mg ha⁻¹. No genotype produced significantly greater DMY than the control; however, four genotypes (Ho 14-9213, Ho 19-9915, Ho 15-9901, and Ho 15-9004) produced significantly less.

Dry matter yields were significantly different among genotypes (p = 0.0010) in the PC year in 2021 (Table 4). The mean DMY for all 20 genotypes was 13.9 Mg ha⁻¹ and the mean for the control was 16.06 Mg ha⁻¹. The genotype with the greatest DMY was Ho 15-9907 (16.95 Mg ha⁻¹), and the genotype with the least DMY was Ho 15-9923 (8.51 Mg ha⁻¹). In 2021, no genotype produced significantly greater DMY than the control, but four genotypes (Ho 15-9905, Ho 15-9915, Ho 15-9918, Ho 15-9923) produced significantly less.

3.3. Extracted Juice Volume

Significant differences in extractable JV among genotypes were observed during the PC year in 2020 (p < 0.0001; Table 3). The mean extractable JV across genotypes was 10.2 kL ha⁻¹ of cane and the mean of the control was 4.9 kL ha⁻¹. The greatest mean JV was 16.9 kL ha⁻¹ (Ho 14-9218) and the least mean JV was 4.9 kL ha⁻¹ (Ho 02-113). Eight other genotypes were the same as Ho 02-113. Twelve genotypes had significantly greater JV than the control.

Extractable JV production was significantly different among genotypes (p < 0.0001) in 2021 (Table 4). The mean JV across genotypes was 24.7 kL ha⁻¹. The genotype with the greatest JV was again, Ho 14-9213, with a mean JV yield of 45.1 kL ha⁻¹. The least mean JV was 14.2 kL ha⁻¹ (Ho 15-9913) which was not significantly different than Ho 02-113 (16.1 kL ha⁻¹). Six genotypes produced significantly greater amounts of JV ha⁻¹ than the control.

3.4. Brix

During the PC year in 2020, there were significant differences in °Brix among genotypes (p < 0.0217; Table 3). When averaged across all genotypes, °Brix was 11.6% and the °Brix of the control variety Ho 02-113 was 11.1%. Variance around the mean of the control prevented separation of any of the genotypes from the control. The maximum and minimum °Brix were 13.7% (Ho 15-9901) and 9.0% (Ho 15-9911), respectively.

Significant differences in °Brix among genotypes (p < 0.0001) were observed in 2021 (Table 4). Across genotypes, mean Brix was 12.6% with a range between 10.2 (Ho 15-9918; the only genotype significantly less than the control) and 15.6 (Ho 15-9901). The control cultivar Ho 02-113 had a °Brix of 12.2 which was significantly less than four other genotypes.

There was a 3% difference in °Brix between years and significant genotypic differences from the control in 2021 only. Mean °Brix values in 2020 ranged from 9.0 to 13.7 (Δ 4.7; Table 3) while in 2021 they ranged from 10.2 to 15.6 (Δ 5.4; Table 4). In 2020, though the range was tighter than 2021, the LSD was 2.6 greater than in 2021 (1.9).

3.5. Theoretical Ethanol from Juice Fermentation

There were significant differences in theoretical ethanol from fermentation (TEF) among genotypes in 2020 (Table 3). The mean TEF among genotypes was 0.5185 kL ha⁻¹ and values ranged from 0.211 kL ha⁻¹ (Ho 15-9904) to 0.791 (Ho 15-9908). The mean TEF of Ho 02-113 was 0.236 kL ha⁻¹ with twelve genotypes producing significantly more ethanol from juice than the control. No genotype produced significantly less TEF than the control.

There were significant differences in theoretical TEF among genotypes in the PC year in 2021 (p < 0.0001; Table 4). The mean TEF among genotypes was 1.35 kL ha⁻¹ with the least yield from Ho 15-9904 (0.806 kL ha⁻¹) and the highest yield from Ho 14-9213 (2.493 kL ha⁻¹). The mean TEF of Ho 02-113 was 0.874 kL ha⁻¹. This site-year produced eight genotypes with significantly greater TEF than the control. No genotype yielded significantly less TEF than Ho 02-113.

It should also be noted that correlations between JV and °Brix (r²₍₈₀₎ = 0.004, p = 0.601317 for 2020 and r²₍₈₂₎ = 0.02, p = 0.28289 for 2021), JV and DMY (r²₍₈₀₎ = 0.006, p = 0.5498 for 2020 and r²₍₈₂₎ = 0.05, p = 0.0437 for 2021) were not significant at α = 0.05. Dry matter yield and TEF was not significantly correlated in either year (r²₍₈₀₎ = 0.003, p = 0.6465 in 2020 and r²₍₈₂₎ = 0.003, p = 0.6458 in 2021).

3.6. Cellulose Percentage

While there were differences in cellulosic percentages among genotypes (p = 0.0198 and 0.0216 for 2020 and 2021, respectively) significant differences were not associated with years (p = 0.2067). The eight data points of cellulose percentage within a genotype (four replications × two years) were averaged and the mean used for subsequent calculations (Table 3 and Table 4). Mean cellulose percentage, presented as a decimal, ranged from 0.34 (Ho 15-9901) to 0.41 (Ho 15-9920). The mean for the control (Ho 02-113) was 0.39. While no genotype had a cellulose percentage greater than the control, three genotypes were significantly less than the control.

3.7. Theoretical Ethanol from Cellulose

There were significant differences in TEC among genotypes in 2020 (p < 0.0025; Table 3). The mean TEC among genotypes was 0.679 kL ha⁻¹ with Ho 02-113 producing 0.779 kL ha⁻¹. The greatest mean TEC was 0.926 kL ha⁻¹ (Ho 15-9918), and the least was 0.408 kL ha⁻¹ (Ho 15-9901). No genotype produced significantly more TEC than the control, but two produced significantly less (Table 3). The genotypes with the least TEC production were Ho 15-9906 (0.440 kL ha⁻¹) and Ho 15-9901.

In 2021 there were significant differences in TEC among genotypes as well (p < 0.0026; Table 4). The mean TEC among genotypes was 0.801 kL ha⁻¹. The genotype Ho 15-9907 produced the greatest TEC (1.007 kL ha⁻¹), and Ho 15-9923 produced the least (0.538 kL ha⁻¹). There were no genotypes with significantly more TEC than Ho 02-113 (0.937 kL ha⁻¹), but there were three genotypes that produced significantly less.

3.8. Total Theoretical Ethanol Yield

Total theoretical ethanol yield (TTEY) was calculated by the addition of TEF to the TEC and reflects the total ethanol that can be produced per hectare when combining production from the fermented juice and cellulosic ethanol. There were significant differences in TTEY among genotypes in 2020 (p = 0.0010; Table 3) and in 2021 (p < 0.00001; Table 4). Mean TTEY among genotypes in 2020 was 1.190 kL ha⁻¹. In 2020, there were four genotypes with mean TTEY greater than the control.

In 2021, mean TTEY among genotypes was 2.158 kL ha⁻¹ (Table 3). The control was grouped among the lowest yielding with a mean TTEY of 1.811 kL ha⁻¹. Three genotypes had significantly greater TTEY; Ho 14-9213 (3.407 kL ha⁻¹), Ho 15-9920 (2.732 kL ha⁻¹), and Ho 15-9921 (2.777 kL ha⁻¹).

Interestingly, the genotype Ho 14-9213 (1.058 kL ha⁻¹) did not differ from the genotype with the least TTEY in 2020 but had the greatest TTEY in 2021 (3.407 kL ha⁻¹).

4. Discussion

4.1. Dry Matter Yield

There was a significant difference in DMY (p < 0.0001) between 2020 and 2021. Mean DMY was significantly greater in 2021 than in 2020 for most genotypes (Figure 1; Table 3 and Table 4). However, there were four genotypes (Ho 15-9905, Ho 15-9913, Ho 15-9918, and Ho 15-9923) that decreased in DMY from the 2020 to the 2021 site-location. The greatest decrease was observed in Ho 15-9923 (−29.0%) and the greatest increase was observed in Ho 15-9909 (+301.2%). As mentioned for plant height, rainfall was more abundant and more uniformly distributed in 2021 [22], and the soil was better drained at the 2021 site-year. Soil temperature in the spring also impacted energycane shoot emergence. In April 2020, during shoot emergence, a cold front delivered 72 mm of rain and dropped the soil temperature at 20 cm from 20° to 15 °C. During the same period in 2021, the soil temperature at 20 cm was 20 °C and on the rise. Temperatures at 15 °C are reported as the minimum required for growth in sugarcane [23]. In 2020, total rainfall was 551.4 mm, while in 2021 cumulative rainfall was 922.5 mm.

Probably the most significant factor influencing DMY (and many of the other characteristics measured) was that the field at the 2021 site-year was fallow for the six years prior to planting this test. Enhanced yields following fallow periods have been previously reported in other crops, thus it was expected that the 2021 site-year would have greater yields as well [24,25,26].

4.2. Juice Volume

Juice volume ha⁻¹ is directly related to total DMY and genotype fiber percentage. Mean JV in 2021 on an area basis was more than double that in 2020 (Table 3 and Table 4). Generally, JV per stalk weight (mL g⁻¹;) is relatively fixed in a genotype and governed by fiber content in the stalk of that genotype [27,28]. As fiber content in the stalk increases, JV decreases, as does the ability to extract the juice with a roller mill. It can, however, be altered by abundant rainfall which saturates the stalk but has a corresponding negative effect on °Brix. The control variety, Ho 02-113, was among the genotypes with the least JV ha⁻¹ and JV (mL g⁻¹).

4.3. °Brix

°Brix is highly subject to the environment. Rainfall can cause °Brix readings to drop substantially within hours [29]. To minimize environmental effects during each year, all °Brix readings were taken during the morning of the same day. We also assumed uniform soil moisture across an entire block, assuming relative values among genotypes to be representative. The differences in the LSD values indicate that values around the means in 2020 were not as tightly clustered as they were in 2021. It should be noted, despite a greater abundance of rainfall in 2021 [22] °Brix readings were greater in 2021 (12.6) than 2020 (11.6), a rough indication of better growing conditions.

4.4. Theoretical Ethanol from Juice Fermentation

A significant difference (p < 0.0001) between TEF in 2020 and 2021 was observed. These differences among genotypes are a complex combination of three factors: JV, which considers stalk fiber, DMY, and to lesser extent °Brix. Juice volume values were based on extraction using a three-roller electric bench top sugarcane juicer. Had this cane been processed using an industrial scale extraction method, the extraction efficiency and juice volume yield would be higher. Mills with higher extraction methods can also affect the Brix level as some moisture within the cane is membrane bound and Brix free. Based on the extraction method used, the mean TEF was significantly greater in 2021 than in 2020. Theoretical ethanol from fermentation of the juice increased among all genotypes in 2021 as compared to 2020. The increase in TEF is not a surprise since all genotypes had a higher JV (and °Brix) in 2021 compared to 2020. In 2020, mean ethanol from fermentation was 44% of the TTEY while in 2021, it was 63%. A definitive statement concerning the contribution of each of the components of TEF seems elusive as none of the components of TEF measured here correlates significantly. As with other characteristics, TEF (juice volume and °Brix) is highly influenced by the environment and extraction method.

4.5. Theoretical Ethanol from Cellulose

Theoretical ethanol yield from corrected DM varied significantly among genotypes and between site-years (p < 0.0001). Like the other characteristics measured, mean TEC values were significantly greater in 2021 than in 2020, although four genotypes produced less TEC in 2020 than in 2021. The relative results are identical between the TEC and the DMY when comparing genotypes. In general, the greater the DMY, the greater is the cellulose per unit area, with the caveat of genotype differences in percentage cellulose. However, correlations run comparing DMY to TEC were strongly positive (r²₍₈₀₎ = 0.95, p < 0.00001 for 2020, and r²₍₈₂₎ = 0.86, p < 0.00001 for 2021).

In the calculation of total theoretical ethanol yield (TTEY), the TEC is the inverse of TEF. The mean TEC for all genotypes in 2020 was 57% and in 2021 mean TEC was 37% of TTEY.

While strongly correlated with DMY, our findings cast doubt on the absolute statement made by Sanford [30] reporting on four species of biomass crops. They indicated that despite differences in sugar production, biomass production was the stronger driver of ethanol production. Energycane genotypes used in this study are considered low sugar by scientists working with Saccharum. Our data suggest we cannot support Sanford’s assertion when the end-goal is ethanol production. Data in this study were influenced by site-year, causing TEC in 2020 to be a greater contributor to TTEY than TEC in 2020 and the opposite in 2021.

4.6. Total Theoretical Ethanol Yield

There was a significant (TTEY) difference among genotypes (p = 0.0096) and between years (p < 0.0001). As with the prior response variables, the mean TTEY value was significantly greater in 2021 than in 2020, owing to the general increased growth rate and DMY in 2021. All genotypes produced less TTEY in 2020 than in 2021. The genotype Ho 14-9213 had a TTEY with a 3.2-fold increase from 2020 to 2021, whereas Ho 15-9918 had yields that were 3% less in 2021. The extreme change in yield of Ho 14-9213 suggests a stronger environmental than genetic component to its TTEY. The inverse is likely true for Ho 15-9918, where yield was virtually stable across site-years. Comparisons of genotypes for TTEY production in 2020 and 2021 show the same five occurred in the top 10 during both years. These genotypes were Ho 15-9907, Ho 15-9918, Ho 15-9919, Ho 15-9921, and Ho 15-9922. Importantly, Ho 15-9921 ranked second in both years and Ho 15-9918, fourth and first in 2020 and 2021, respectively. The ability to maintain rank position, especially in a changing environment, is an important indication of the genetic component’s substantial and overriding influence on phenotype. In a breeding sense, these data suggest that more progeny from the parents of Ho 15-9921 and Ho 15-9918 should be evaluated for their genetic contribution to future genotypes to be tested in northcentral Mississippi.

4.7. Correlations with Total Theoretical Ethanol Yield

To determine if DMY could be used to predict TTEY, correlations were made between traits for each site-year in the study. Even though both r² values were significant, their predictive value was moderate to low (

Table 5. Correlation table for dry matter yield (DMY), theoretical ethanol yield from fermentation (TEF), theoretical ethanol yield from cellulose * (TEC), and percentage cellulose with total theoretical ethanol yield (TTEY).

	DMY		TEF		TEC		%Cellulose
Year	2020	2021	2020	2021	2020	2021	2020	2021
TTEY	0.46	0.37	0.53	0.91	0.54	0.34	0.34	0.11
	***	***	***	***	***	***	***	NS

* To correct for variance due to soluble solids, total cellulose yield was calculated as percentage cellulose × corrected DMY (field dry weight minus soluble solids). *** indicates correlation highly significant (α = 0.001). NS indicates correlation was not significant at α = 0.05.

). The coefficient of determination in 2020 was r²₍₈₀₎ = 0.46 (p < 0.00001) and in 2021 r²₍₈₂₎ = 0.37 (p < 0.00001). Correlations of percentage cellulose and TTEY (r²₍₈₀₎ = 0.3403, p = 0.002011) in 2020 were significant, but not in 2021 (r²₍₈₂₎ = 0.1146, p = 0.30527 in 2021) making percentage cellulose an unlikely value to predict TTEY. To correct for variance due to soluble solids, total cellulose yield was calculated (percentage cellulose × corrected DMY); however, the coefficient of determination for 2020 was not significant (r²₍₈₀₎ = 0.01, p = 0.2931) while in 2021 it was moderately predictive and significant (r²₍₈₂₎ = 0.33, p < 0.00001.

Fermented juice and cellulose components when added together give TTEY. Both components were significantly correlated with TTEY for both years. Total ethanol from fermentation and TTEY correlations were (r²₍₈₀₎ = 0.53, p < 0.00001 in 2020 and r²₍₈₂₎ = 0.91, p < 0.00001 in 2021). Total ethanol from fermentation was strongly correlated with TTEY in 2021, indicating for every 56-increment increase in ethanol produced from fermentation, there was a nearly equal increase in TTEY, while TEC and TTEY correlations (r²₍₈₀₎ = 0.54, p < 0.00001 in 2020 and r²₍₈₂₎ = 0.34, p < 0.00001 in 2021) were less strongly correlated.

5. Summary and Conclusions

This test location was 345 km north of normal Saccharum production areas. Over the two years of this test, Ho 02-113 was exceeded in mean DMY by only Ho 14-9213. When considering mean TTEY, 16 of the 19 test genotypes equaled or exceeded Ho 02-113. Theoretical ethanol from fermentation as a component of TTEY was highly correlated in 2021; however, in 2020, TEF and TEC and in 2021 were only moderately correlated to TTEY. The lack of strong correlations is likely due to the change in contribution ratio of TEF and TEC over the two site-years of this study. The percentage of TEF was 43% and 63% in 2020 and 2021, respectively, indicating that this fraction of energycane yield, even when aiming toward cellulosic conversion for ethanol production, should not be discounted. The genotypes tested in this study were considered low-sugar, yet fermentable sugars were still a substantial contributor to TTEY estimates.

When we look specifically at the difference in TTEY for Ho 15-9918 between 2020 and 2021, it was roughly three percent, indicating yield stability across the two very different site-years. This suggests the genetics of specific genotypes were able to overcome the environmental effects between the site-years.

Although Ho 15-9903 was not among the top 10 in 2021, it is important to note that its mean total DMY between the two site-years during the PC production were within two percent of each other. This consistency, in spite of very different environments, indicates genetic stability. Other yield-stable genotypes include Ho 15-9912, Ho 15-9925, and Ho 02-113, which all yielded among the top 10 during both site-years. However, yield stability is not the final goal. Genotypes with consistently poor yield and ethanol output are not a benefit to the energycane breeding program and testing such as this gives justification for dropping these genotypes (or their parents) from the program. Additional, actual on-site testing is necessary to identify genotypes that are suited for biomass production in areas north of the traditional Saccharum cultivation locations.