Managing flowering time in Miscanthus and sugarcane to facilitate intra- and intergeneric crosses

Miscanthus is a close relative of Saccharum and a potentially valuable genetic resource for improving sugarcane. Differences in flowering time within and between Miscanthus and Saccharum hinders intra- and interspecific hybridizations. A series of greenhouse experiments were conducted over three years to determine how to synchronize flowering time of Saccharum and Miscanthus genotypes. We found that day length was an important factor influencing when Miscanthus and Saccharum flowered. Sugarcane could be induced to flower in a central Illinois greenhouse using supplemental lighting to reduce the rate at which days shortened during the autumn and winter to 1 min d-1, which allowed us to synchronize the flowering of some sugarcane genotypes with Miscanthus genotypes primarily from low latitudes. In a complementary growth chamber experiment, we evaluated 33 Miscanthus genotypes, including 28 M. sinensis, 2 M. floridulus, and 3 M. ×giganteus collected from 20.9° S to 44.9° N for response to three day lengths (10 h, 12.5 h, and 15 h). High latitude-adapted M. sinensis flowered mainly under 15 h days, but unexpectedly, short days resulted in short, stocky plants that did not flower; in some cases, flag leaves developed under short days but heading did not occur. In contrast, for M. sinensis and M. floridulus from low latitudes, shorter day lengths typically resulted in earlier flowering, and for some low latitude genotypes, 15 h days resulted in no flowering. However, the highest ratio of reproductive shoots to total number of culms was typically observed for 12.5 h or 15 h days. Latitude of origin was significantly associated with culm length, and the shorter the days, the stronger the relationship. Nearly all entries achieved maximal culm length under the 15 h treatment, but the nearer to the equator an accession originated, the less of a difference in culm length between the short-day treatments and the 15 h day treatment. Under short days, short culms for high-latitude accessions was achieved by different physiological mechanisms for M. sinensis genetic groups from the mainland in comparison to those from Japan; for mainland accessions, the mechanism was reduced internode length, whereas for Japanese accessions the phyllochron under short days was greater than under long days. Thus, for M. sinensis, short days typically hastened floral induction, consistent with the expectations for a facultative short-day plant. However, for high latitude accessions of M. sinensis, days less than 12.5 h also signaled that plants should prepare for winter by producing many short culms with limited elongation and development; moreover, this response was also epistatic to flowering. Thus, to flower M. sinensis that originates from high latitudes synchronously with sugarcane, the former needs day lengths >12.5 h (perhaps as high as 15 h), whereas that the latter needs day lengths <12.5 h.


Introduction
Miscanthus is an emerging bioenergy biomass crop in North America and Europe [1,2]. As a C 4 perennial grass, Miscanthus is native to eastern Asia and Oceania from tropical to cold-temperate environments [3]. However, currently only one single triploid clone of M. ×giganteus, which is an interspecific hybrid between M. sinensis and M. sacchariflorus, is widely available for commercial production and new hybrids are needed. Additionally, Miscanthus is a close relative to Saccharum and is potentially a valuable genetic resource for improving sugarcane [4][5][6][7][8].
Control of flowering time is important to plant breeders because it allows them to make crosses of their choosing. Constraints on which genotypes can be used as parents in crosses would be severe impediments to plant improvement. Synchronization of flowering time between sugarcane and Miscanthus is necessary for making intergeneric crosses between these two species, because, like most warm-season grasses, pollen of these two genera quickly loses viability within the first~2 h of dehiscence under typical growing conditions [9][10][11][12]. Moreover, because Saccharum and Miscanthus pollen is typically intolerant of desiccation, it is not readily stored frozen; thus, consistently effective and long-term pollen-storage methods have not yet been developed for these genera.
M. sacchariflorus has been considered a quantitative short-day plant [13], similar to sorghum and sugarcane. M. sinensis was described as day neutral by Deuter [14], whereas Jensen et al. [15] reported that flowering time in M. sinensis was more complicated, depending on multiple factors, including thermal time, temperature, photoperiod, and precipitation. In the field at Urbana, M. sacchariflorus flowers as early as July and as late as early November, whereas M. sinensis flowers from late July to mid-October [16,17]. In sugarcane, floral initiation is induced by a small decrease (30-60 sec per day) in day length from about 12.5 h [18,19]. Most sugarcane varieties need between 12 and 12.5 h of photoperiod to induce flowering [20][21][22]. In our greenhouses at Urbana, Illinois, flowering of diverse Miscanthus accessions typically is greatest from August through October and again from March through June. For most sugarcane breeding programs in the U.S., peak flowering is in November and December. In central Illinois, the rapid decrease in day length during the autumn is not conducive to flowering sugarcane plants in the greenhouse. Thus, it would be desirable to develop methods to synchronize the flowering time of Miscanthus and Saccharum, thereby facilitating the introgression of desirable genes for improving sugarcane. Additionally, it would be advantageous to be able to better predict and control flowering time in Miscanthus so that we can more readily make crosses between different Miscanthus genotypes.
In this study, we conducted one set of experiments to explore the feasibility of synchronizing flowering time of Saccharum and Miscanthus in a central Illinois (~40˚N) greenhouse, and a complementary experiment in growth chambers to understanding how day length impacts flowering time and plant growth of M. sinensis. The objectives were: 1) to assess the diversity of flowering time for Miscanthus and Saccharum, 2) to determine the effects of cultural treatments that we hypothesized could delay flowering time in Miscanthus, 3) to determine how day length in controlled environment chambers affects flowering time of M. sinensis accessions that originate from different latitudes.

Experiment 1: Flowering time management of Miscanthus and sugarcane in a greenhouse
To determine how to synchronize the flowering of Miscanthus and Saccharum, a series of greenhouse experiments were conducted over three years (2014-2017; Expts. 1a-c). A key component of the study was to assess the diversity of flowering times within each genus, when plants were grown in a greenhouse at Urbana, IL under a photoperiod treatment that was expected to be conducive to flowering of sugarcane. We also evaluated cultural treatments that we hypothesized had the potential to delay flowering of Miscanthus, such as 4˚C cold storage to delay the start of growth, cutting plants to 15 cm above the soil surface, and the combination of cutting followed by one month of 4˚C cold storage.
A panel of 23 Miscanthus (Table 1) and 31 Saccharum accessions (Table 2) were studied. All plants were grown in a tall (6.1 m eave height), controlled-environment greenhouse at the University of Illinois Energy Farm at Urbana, IL (40.1˚N, 88.2˚W), located where there was no light pollution (e.g. from street lamps or buildings) that could interfere with the short-day treatment required to flower sugarcane. When natural day length reached 12.5 h in Urbana (14 September in 2014(14 September in , 2015(14 September in , 2016, supplemental light (MH 1000W/U/BT37 metal halide bulbs, Venture Lighting, Twinsburg, OH, US) was provided to decrease the day length by 1 min d -1 until a photoperiod of 11 h d -1 was reached (13 December in 2014(13 December in , 2015(13 December in , 2016, at which point the day length was held constant until exceeded by the natural day length (22 February in 2014(22 February in , 2015(22 February in , 2016. Additionally, in the third year experiment (2016-2017), we grew a second set of the Miscanthus genotypes in a greenhouse on the University of Illinois main campus (<5 km from the Energy Farm greenhouse), in which the plants were given constant 13 h d -1 photoperiod, starting on 2 September until natural day length exceeded this value on 9 April. In the greenhouses, temperature during the day was maintained between 27-31˚C and at night temperature was between 22-26˚C. Miscanthus plants were grown in 7 L pots (T.O. Plastics, Clearwater, MN, USA) containing peat-based potting mix (Metro-Mix 900, Sun Gro Horticulture, Agawam, MA, USA), whereas the larger-growing Saccharum plants were grown in 17 L pots. Slow release fertilizer was applied to each pot (Osmocote Pro 17-5-11, 6 months; 35 g per 7 L pot and 140 g per 17 L pot; ICL Specialty Fertilizers, Dublin, OH, USA). Drip irrigation was supplied to each pot automatically twice per day. For each pot of Miscanthus and Saccharum studied, data was recorded weekly when a plant was actively flowering (newly opened florets dehiscing pollen).
The   (Table 1). On 6 September, an initial set of six stored divisions per Miscanthus genotype were removed from cold storage and three were planted in the greenhouse running the 1 min d -1 decreasing photoperiod protocol and another three divisions were planted in another greenhouse with a constant 13 h d -1 day length. In total, three sets of 4 C Miscanthus divisions were planted in each greenhouse at 4-week intervals from September to November. Data on flowering time was recorded weekly from 1 October 2016 to 30 April 2017. In total, 33 Miscanthus genotypes and two Sorghum bicolor controls (one short-day and one day-neutral) were studied ( Table 3). The Miscanthus genotypes included 25 M. sinensis from known locations in China and Japan, representing latitudes ranging from 19 to 45˚N, three ornamental M. sinensis cultivars, two M. floridulus from New Guinea and New Caledonia, two diploid M. ×giganteus (one ornamental cultivar and one natural hybrid), and the leading biomass cultivar control, the triploid M. ×giganteus '1993-1780'. The M. sinensis genotypes studied here represent six genetic groups that were previously identified by Clark et al. [23,24]. Although detailed source location information for the four ornamental cultivars and the M. ×giganteus '1993-1780' control is not available, their M. sinensis ancestry was previously shown to be from the Southern Japan genetic group (Table 3 [23,24]).
Plants were established in 7 L pots in controlled environment chambers under constant long days (15 h). After 42-61 d of establishment in the chamber, all the aboveground stems of the Miscanthus plants were cut to 5 cm above the soil surface and then subjected to one of three day length treatments: 15 h, 12.5 h, and 10 h. For each combination of genotype and day length treatment, three replicate pots were tested. The temperature was a constant 23˚C for the duration of the experiment. To each pot, 35 g of slow release fertilizer (Osmocote Pro 17- 5-11, 6 months; ICL Specialty Fertilizers, Dublin, OH, USA) was added at planting and after 6 months. Drip irrigation was provided to each pot. Data were recorded on the number of days to first flagging and first flowering. At the end of the experiment, data were taken on number of total culms and number of reproductive shoots, number of leaves per culm (~number of nodes), and culm length. An additional trait, reproductive shoot ratio, was obtained by dividing number of reproductive shoots over the total culm count. Thus, a total of seven traits were studied. The experiments were ended after at least 80 d with no change in flowering, which was at least 188 d from cutting for the 10 h and 12.5 h treatments and 352 d for the 15 h treatment.

Statistical analysis
For Experiment 1, analyses of variance (ANOVAs) were conducted to assess the effects on Miscanthus flowering time of the treatments performed in each year. For the 2014-2015 experiment, the treatments included cut, and cut plus cold performed monthly from September to January and controls. For the 2015-2016 experiment, the treatments were plantings of pot divisions or rhizomes from cold storage, performed monthly from June to September, and controls. For the 2016-2017 experiment, the treatments were plantings of cold storage pot divisions from September to November, grown under two different day lengths, and controls. ANOVAs were conducted with SAS Procedure MIXED (SAS Institute Inc., Cary, NC, USA) for each year's experiment based on the subset of Miscanthus genotypes that flowered following the model: where Y is first flowering time, T represents treatment, G equals genotype, M represents month, R represents replication, and TG, GM, TM, TGM represent respective interactions of aforementioned model terms, and ε is error. Treatment, genotype and month were set as fixed and replication was set as random. To better evaluate flowering time diversity between and within Miscanthus and Saccharum, ANOVAs were also conducted in SAS Procedure MIXED to test the effects on flowering-time of genus (Miscanthus, Saccharum), and genotype nested within genus as fixed effects, for the subset of genotypes that flowered; for Miscanthus, only the control pots were included in this analysis. Weekly flowering data were plotted in R [25] for visualization. Association between the latitude of origin for the Miscanthus genotypes and flowering time was also evaluated by linear regression using R lm function [25].
For Experiment 2, ANOVAs were conducted with SAS Procedure MIXED to assess the fixed effects of genotype, day length (10 h, 12.5 h, and 15 h) and their interactions on flowering traits (days to first flagging and first flowering) and morphological traits (culm length, number of leaves per culm, number of total culms, number of reproductive shoots and reproductive shoot ratio). Tukey's HSD test (α = 0.05) was estimated to investigate differences among three day lengths for each trait. The relationships between location of origin (i.e. collection site), the genetic groups to which the genotypes belong, and the phenotypic traits observed in the controlled-environment chambers under three day lengths were visualized using R package ggmap [26] by plotting on a geographical map the location of each genotype, color coded by its previously ascertained M. sinensis genetic group [23,24], along with the associated phenotypic data from this study (as bar charts with standard errors). Associations between the latitude of origin and phenotype at the different day lengths were also evaluated by linear regression using R lm function [25]. R codes used in figure visualization are available at https://github.com/hxdonggenetics/Geographic-map-in-Miscanthus-flowering-study.

Experiment 1: Flowering time management of Miscanthus and sugarcane in a greenhouse
Key findings over the three years. Large and highly significant differences in flowering time were observed between Saccharum and Miscanthus, and among genotypes within each genus (Fig 1; Table 4). As expected Saccharum genotypes typically flowered later than Miscanthus genotypes. However, some Saccharum and Miscanthus genotypes overlapped in flowering time each year the experiment was conducted (Fig 1). Each year, the experiment was initiated~2 months later in the season than the prior year (Expt. 1a, 25 Fig 2). Thus, under the short days provided, Miscanthus genotypes that originated from low latitudes were primarily the ones that overlapped in flowering time with Saccharum genotypes (Figs 1 and 2).
Some Miscanthus and Saccharum genotypes flowered consistently over the three years that the experiment was conducted, irrespective of the differences in initial planting date.  Table). One Saccharum hybrid, 'Ho91-552', flowered a second time in January 2015 and three Saccharum hybrids, 'L09-105', 'L79-1002' and 'Ho06-9002', had second flush of flowering in April 2015 (Fig 1).
For Miscanthus, the control pots of the 2014-2015 experiment flowered only from August through December (Fig 1, S1 Table). The earliest flowering genotype was the northernmost M.  Table). ANOVA indicated that genotype, treatment, month of treatment application, and interactions all had significant effects on days to    (Fig 1, S2 Table).
Miscanthus pot divisions and rhizomes that were stored at 4˚C then planted in the greenhouse during June or July flowered in high frequency, but few or no genotypes flowered when cold-stored materials were planted in August or September, again demonstrating that date of establishment had a large effect on presence or absence of flowering (Fig 1, S2 Table). However, flowering time of the cold-stored Miscanthus divisions and rhizomes was similar to that of the controls. ANOVAs indicated that all tested model terms had significant effects except for treatment by month interaction and genotype by treatment by month interaction (Table 5). Of the 15 M. sacchariflorus genotypes included in the treatments, 11 flowered from stored pot divisions (seven each from June and July plantings but only one from August and zero from September; S2 Table), and all flowered when pots were newly established from rhizomes (15 from June, 12 from July, one from August, and zero from September; S2 Table). For the Miscanthus, only 10 of the 23 genotypes flowered, and of these, two flowered only after cold-stored divisions were planted in September or October (Fig 1, S3 Table). However, of the 15 Miscanthus genotypes included in the cold storage treatments, only four flowered (Fig 1, S3 Table). An ANOVA of just the four entries that flowered to evaluate effects of genotype, two day length treatments, month and their interactions on days to first flowering, detected significant effects of genotype and day length ( Table 5

Experiment 2: Effect of day length on flowering time of M. sinensis, M. floridulus, and M. ×giganteus '1993-1780' in controlled environment chambers
ANOVAs indicated that genotype, day length, and genotype by day length interactions had significant effects on each of the seven flowering and morphological traits ( Table 6). All 35 In   (Fig 3, Table 3). Similarly, for the Miscanthus genotypes that flowered under 10 h and 12.5 h, average days to first flower (64 and 90 d, respectively; Table 3) were earlier than those that flowered at 15 h (151 d), though the difference between 10 h and 12.5 h was not significant at α = 0.05 based on Tukey's HSD test (Fig  3). The day-neutral S. bicolor control '38M' (ma 1 ma 2 ma 3 R Ma 4 ; [27,28]) flowered quickly and at about the same time regardless of day length (50 to 60 days after cutting), as expected; however, none of the Miscanthus genotypes behaved similarly (Fig 3, Table 3).
Of the 33 Miscanthus genotypes, all but three tropical accessions flowered under the 15 h day length (Fig 3, Table 3), and the highest ratio of reproductive shoots to total number of culms was typically observed for 15 h days (Fig 4B, S4 Table). With the 15 h day length, days to first flower for the M. sinensis genotypes ranged from 66 d to 360 d (Table 3) (Fig 3, Table 3). Similarly, for four of the five tropical Miscanthus genotypes, reproductive shoot ratio was highest under 12.5 h days, in contrast to those that originated at higher latitudes (Fig 4B, S4 Table).
At 10 h day length, there was a strong negative correlation between the latitude of origin and days to first flower (r 2 = 0.88), but at 12.5 and 15 h, the correlations were only moderately negative (Fig 3). However, none of the 12 M. sinensis genotypes that originated from latitudes exceeding 34˚N flowered under 10 h days, and only one ('EBI-2008-051c') of these flowered under 12.5 h days, yet all flowered under 15 h days (Fig 3, Table 3). Notably, six of these northern (i.e. temperate) M. sinensis genotypes flagged under 10 h and/or 12.5 h day lengths but did not proceed to flower (Fig 3; ' h days or all three tested day lengths, indicating that the subtropics is a transition zone with a mixture of day length response types (Fig 3). Moreover, in addition to not flowering under short days, the northern M. sinensis genotypes responded to 10 and 12.5 h days by producing very short culms, with the shortest days resulting in the shortest culms (Figs 5 and 6, S4 Table).
Culm length of the M. sinensis and M. floridulus genotypes was strongly and negatively correlated with latitude of origin under 10 h days (r 2 = 0.81) and 12.5 h days (r 2 = 0.63), but the relationship was weak under 15 h days (r 2 = 0.09; Fig 5A). Among all 33 Miscanthus genotypes, Tukey's HSD test (α = 0.05) indicated that culm length was significantly different across three day length treatments. Nearly all the Miscanthus entries achieved maximal culm length under the 15 h treatment (including the biomass cultivar M×g '1993-1780'), but the nearer to the equator an accession originated, the less of a difference in culm length between the short day treatments and the 15 h day treatment (Fig 5A). For example, M. floridulus 'NG77-022' from 3.6˚S produced similarly long culms under all three day lengths (Fig 5A and S1 Fig, S4 Table). Two tropical genotypes ('PMS-382' and 'US56-0022-03'), two subtropical genotypes ('PMS-226' and 'Miyazaki') and one ornamental cultivar ('Cabaret') were tallest under 12.5 h days (Fig 5A, S4 Table).
M. sinensis genotypes that originated from high latitudes in Japan had greater numbers of leaves at 15 h than at 10 h day lengths (Fig 5B, S4 Table). In contrast, M. sinensis genotypes that originated from high latitudes on mainland Asia (Korea/North China and Yangtze-Qinling genetic groups) had the same or greater numbers of leaves at 10 h in comparison to 15 h (Fig 5B, S4 Table). Thus, for the Japanese accessions, the short culms observed for high-latitude accessions of M. sinensis under short days was achieved substantially by greater phyllochron under short days than under long days, whereas for the mainland accessions, short culms were obtained primarily via short internodes rather than by more days needed to develop a leaf. Like the northern Japanese M. sinensis genotypes, most of the subtropical and tropical accessions of M. sinensis produced more leaves under long days than under short days. However, some accessions produced similar numbers of leaves under all three day lengths tested (e.g. 'PMS-306', 29.9˚N), and other entries, such as M. floridulus 'NG77-022' (3.6˚S) and the biomass control cultivar M×g '1993-1780' produced more leaves under shorter days than longer days (Fig 5B, S4 Table). Total number of culms for most of the Miscanthus genotypes was~3-13 fold greater under 10 h than 15 h days, with intermediate numbers of culms typically resulting from 12.5 h days (Fig 4A, S4 Table). However, the two tropical M. floridulus ('NG77-022' and 'US56-0022-03'), four M. sinensis ('Flamingo', 'Koike-21c', 'Miyazaki', and 'Tohoku-2010-015a'), and the biomass control M×g '1993-1780' produced the greatest number of culms at 12.5 h. Thus, under 10 and 12.5 h day lengths, most of the M. sinensis genotypes from low latitudes produced a large number of tall culms, many of which flowered, whereas genotypes from high latitudes produced a large number of short culms that did not flower (Figs 4-6, S1 and S2 Figs).
Raw data for experiment 1 and experiment 2 can be found in S5 and S6 Tables, respectively.

Flowering sugarcane at 40˚N
Flowering was accomplished for more than half of the sugarcane genotypes in this study, in central Illinois, by growing the plants in a warm greenhouse and providing a declining   photoperiod of 1 min d -1 from 12.5 h to 11 h over the course of 3 months, then holding a constant 11 h day length for an additional~2 months. Sugarcane is difficult to flower and synchronize for crosses, so sugarcane breeders commonly use photoperiod facilities to induce flowering by an initial exposure to~12.5 days followed by a declining day length of 30-60 sec d -1 [18,19,29,30]. Further improvements in the number of genotypes that can be flowered in our greenhouse might be obtained by adjusting the rate of decline in photoperiod. Recently, two studies found that a photoperiod decline of 40-45 sec d -1 was likely superior to 30 or 60 d -1 for flowering most sugarcane genotypes [31,32].
The early establishment of the sugarcane pots in Expt. 1a relative to Expts. 1b and 1c was advantageous, resulting in more than twice as many genotypes flowering in autumn and early winter, and also enabling a second flush of flowering for some genotypes in late winter and spring that was not obtained in the later-planted experiments. Julian et al. [33] and Berding [19] observed that the optimal age of sugarcane stems for floral induction was 12-16 weeks. In our study, when the critical 12.5 h photoperiod was reached in mid-September, the age of the from China and Japan both had short culms when grown under short days, indicating different mechanisms of responding to day length resulting in similar height outcomes.
https://doi.org/10.1371/journal.pone.0240390.g005 sugarcanes was~20 weeks for Expt. 1a,~14 weeks for Expt. 1b, and 6 weeks for Expt. 1c. Thus, under our conditions, an establishment phase about six weeks longer than the~14 weeks optimum previously reported was beneficial. Though the later planting of sugarcane in Expts. 1b and 1c helped limit height, thereby avoiding stems reaching the roof of a greenhouse with 6.1 m side-walls, the height problem could be better addressed by air layering stems so that they could be cut if they get too tall, without sacrificing growth. Air layering would also make it easier for workers to move stems during flowering to facilitate emasculation and crossing.
Species and genotype also had a large effect on timing and ease of flowering of sugarcane in our study. The earliest flowering species were S. spontaneum and S. arundinaceum, which was expected [34]. Saccharum hybrids with a high proportion ancestry from S. spontaneum, such as 'L79-1002', 'Ho06-9001', and 'Ho06-9002', were among the most consistent to flower in our study. However, some commercial sugarcane materials, such as 'L09-105', also flowered well in our study.

Effects of day length on Miscanthus development
Photoperiod profoundly affected all aspects of Miscanthus growth and development that we studied, especially flowering. Expt. 2 demonstrated that few M. sinensis or M. floridulus genotypes that originated outside of the tropics flowered well under 12.5 h days or less, yet all the subtropical and temperate-sourced genotypes flowered well under 15 h days (Fig 3), which is the photoperiod during the summer solstice at 40˚N, where Urbana is located. Jensen et al. [13] concluded that M. sacchariflorus is a quantitative short-day plant because flowering under a constant 12.5 h or a declining photoperiod mimicking 34.1˚N was >50 days earlier than for those grown under constant 15.3 h days, which was generally consistent with our observations for M. sinensis in Expt. 2, though critical photoperiods may vary by species and genotype. For M. sacchariflorus grown under a declining photoperiod mimicking 34.1˚N, Jensen et al. [13] estimated that floral induction occurred between 13.8 and 12.5 h day lengths.
Notably, Jensen et al. [13] also observed that M. sacchariflorus genotypes originating from 34.5˚N and higher failed to flower under a declining photoperiod mimicking 34.1˚N, even though some produced flag leaves when day lengths were between 12.7 and 12.1 h; in contrast, M. sacchariflorus genotypes from lower latitudes flowered when days were shorter than 12 h. For M. sinensis, we similarly observed that flowering of genotypes from temperate latitudes (>34˚N) was inhibited by short days (constant 10 and 12.5 h), even though some produced flag leaves, whereas flowering was consistently strong under 15 h days. In addition to not flowering, M. sinensis from temperate latitudes produced many short culms under 10 and 12.5 h days, resulting in a short and dense morphology similar to that of many alpine plants (Figs 5 and 6, S1 and S2 Figs). Such a dense and short morphology can protect apical meristems from freeze damage by keeping them below the soil surface, and limit water loss by reducing air flow around leaves. Thus, for Miscanthus, relatively short days can accelerate floral induction, but below a critical threshold, especially for genotypes adapted to high latitudes, short days can signal that plants should prepare for winter, and importantly this response is epistatic to flowering. Similarly, short-days (<12.5 h) have been shown to induce dormancy and reduce or prevent flowering in switchgrass (Panicum virgatum) and big bluestem (Andropogon gerardii) (especially for high-latitude populations), which are also quantitative short-day, perennial, C 4 grasses [35][36][37]. Moreover, low-intensity light extension of day length prevented or reversed this dormancy in switchgrass [38].
In the greenhouse experiment (Expt. 1), we established Miscanthus plants at different times (implemented by different initial planting dates, by cutting back established plants, or by cutting back plants then storing them at 4˚C for 1 month to mimic dormancy) in an effort to identify treatments that could delay flowering sufficiently to synchronize with sugarcane, but time of establishment was only effective if day length was conducive. Establishing Miscanthus plants from March through the first week of July enabled genotypes from subtropical and temperate latitudes to flower in late summer and early autumn (Fig 1; Expts. 1a and 1b), indicating that floral induction occurred during photoperiods greater than 12.5 h, prior to mid-September, which was consistent with the results of Expt. 2 and Jensen et al. [13]. Moreover, there was little difference in flowering time between plants of the same genotype established in June compared to those established in early July (Fig 1; Expt. 1b), indicating that more rapid flowering associated with the shorter photoperiods encountered by mature stems of the later planting compensated for the difference in planting date. Thus, when established in spring and early summer, the Miscanthus genotypes from subtropical and temperate latitudes flowered early and failed to synchronize with most of the sugarcane genotypes, though some overlap was achieved with the early-flowering S. spontaneum and S. arundinaceum accessions. With early-season establishment and under the declining photoperiod treatment during autumn in the greenhouse, only the two tropical Miscanthus genotypes tested (M. floridulus 'US56-002-03' and M. sinensis 'PMS-375') flowered late enough to consistently synchronize flowering with the first flush of sugarcane flowering in Expt. 1a (in late November and early December) and the single flush of sugarcane flowering in Expts. 1b and 1c (Fig 1), which was consistent with the results of Expt. 2 that these low-latitude genotypes flowered strongly under constant 12.5 h days but did not flower under 15 days (Fig 3). When Miscanthus genotypes from subtropical and temperate latitudes were established during the last week of July or later in the summer or autumn, few flowered because the photoperiod was too short to be conducive by the time stems had sufficiently matured; the exceptions were primarily M. sacchariflorus genotypes, and the tropical M. floridulus 'US56-002-03' and M. sinensis 'PMS-375' (Fig 1; Expts. 1a-c). For example, when some M. sacchariflorus genotypes were established during the first week of September, flowering was delayed until November, which would allow synchronization with many sugarcane genotypes (Fig 1; Expt. 1c).

Synchronizing flowering time of sugarcane and Miscanthus to facilitate intergeneric crosses
To synchronize flowering of sugarcane and Miscanthus in the autumn, it would be advantageous to hasten flowering of the sugarcane and delay flowering of the Miscanthus. Furthermore, it would be desirable to promote flowering of both genera during the late winter and spring. To achieve strong flowering of sugarcane, in a high-latitude greenhouse such as ours, during autumn and early winter, and promote flowering in spring, the plants should be established from cuttings five to six months prior to onset of the 12.5 h and declining day lengths critical for floral induction.
For Miscanthus that originated from the tropics, the same environment that is conducive to flowering of sugarcane, including declining photoperiod, will likely result in synchronized flowering between the two genera during the late autumn. Moreover, cutting back established plants of tropical Miscanthus genotypes in early September, December or January can be used to delay flowering and synchronize with a second spring flush of sugarcane flowering. We note, however, that cold treatments after cutting were disadvantageous for flowering tropical Miscanthus genotypes.
For M. sinensis genotypes that originated from subtropical and temperate latitudes, however, the short and declining day lengths needed to flower sugarcane are not conducive to synchronization of flowering between the two genera. One strategy for synchronizing the flowering of subtropical and temperate M. sinensis is to grow the plants under a conducive photoperiod, such as constant 15 h days (in controlled environment chambers or in a different greenhouse than that used to grow the sugarcanes) and use empirical data on the number of growing days needed to obtain first or peak flowering (e.g. S1-S3 Tables) to choose a planting date that would achieve concurrent flowering with sugarcane in late autumn and early winter or in spring. Though data from Expt. 2 indicated that a constant 15 h day length should facilitate strong flowering after a defined number of days for most if not all subtropical and temperate M. sinensis, it may not be the fastest or optimal day length. Given that 12.5 h days was observed to be too short, an optimal day length for flowering subtropical and temperate M. sinensis may be between 13 and 15 h, though further testing would be needed to determine this. Moreover, Castro et al. [37] found that providing switchgrass, a cumulative short-day plant, with 24 h photoperiod, resulted in multiple rounds of flowering and this could be used to synchronize flowering between early and late genotypes. Given these promising results from switchgrass and the high level of flowering observed under~15 h days in M. sinensis (Expt. 2) and M. sacchariflorus [13], it would be worthwhile to investigate if a 24 h photoperiod would also produce sequential flowering in Miscanthus.
For M. sacchariflorus grown under the short and declining photoperiod needed to flower sugarcane, most genotypes flowered as late as the end of October, which was still too early to synchronize with most sugarcane genotypes. However, M. sacchariflorus ssp. lutarioriparius 'PF30022' was a notable exception, in that plants given a cut plus 1 month cold treatment in September, December or January then grown under the short and declining day length regime that was conducive to flowering sugarcane, produced flowers in late November or March/ April, which would match well with sugarcane flowering (Fig 1; Expt. 1a). M. sacchariflorus ssp. lutarioriparius is indigenous to the lower Yangtze River watershed and is a tall plant with high-biomass yield that is harvested locally to produce paper on a commercial scale [3,39,40], so crossing it to sugarcane would be desirable. However, to delay flowering of most M. sacchariflorus genotypes for synchronization with sugarcane, we suggest cultivation of the former under a constant conducive photoperiod for an empirically determined amount of time, similar to the strategy we propose for subtropical and temperate M. sinensis. However, there is currently little information on what might be optimal photoperiods for flowering M. sacchariflorus. Jensen et al. [13] observed that M. sacchariflorus flowered under constant 15.3 h days, so that would be one option. We observed that under constant 13 h days, three out of six M. sacchariflorus genotypes from eastern Russia planted during the first week of October began to flower by early December (Fig 1, Expt. 1c), which would be suitably late for crossing with sugarcane; however, because these accessions originated from~49˚N, an optimal day length for flowering them might be expected to be greater than 13 h. Given that M. sacchariflorus originates from a wide range of latitudes, day lengths that are optimal for flowering might be expected to range from 12.5 to 16 h.
In this study, we identified barriers to synchronizing the flowering of sugarcane and Miscanthus, and proposed methods to circumvent these. For a given genotype of Miscanthus, a range of flowering dates may be obtained by staggered plantings grown under a single conducive constant day length, or by planting on a single date and growing under a range of conducive and constant day lengths, leveraging the short-day response of faster flowering under shorter day lengths than longer ones. By controlling flowering time of sugarcane and Miscanthus, plant breeders will be better able to improve these crops via intra-and intergeneric crosses of their choosing. Miscanthus genotype names represent the M. sinensis genetic groups identified by Clark et al. [23,24], which included Korea/North China (red), Yangtze-Qinling (green), Northern Japan (blue), Southern Japan (yellow), Sichuan Basin (orange), and Southeastern China plus tropical (purple). In each photo, plant size is scaled by either a 20 cm ruler (black and white) or a 1 m stick (orange and white).  [23,24], which included Korea/North China (red), Yangtze-Qinling (green), Northern Japan (blue), Southern Japan (yellow), Sichuan Basin (orange), and Southeastern China plus tropical (purple); for interspecific hybrids (PMS-300) between M. sacchariflorus and M. sinensis, the dominant M. sinensis genetic group is shown. In each photo, plant size is scaled by either a 20 cm ruler (black and white) or a 1 m stick (orange and white). Note that accessions originating from high latitudes typically remained short and had few or no flowering stems when grown under short days but were taller and flowered when grown under long days. (TIF) S1 Table