Assessment of Alfalfa Populations for Forage Productivity and Seed Yield Potential under a Multi-Year Field Trial

Abstract

Alfalfa is the most important forage legume in the production of voluminous fodder. Although not primarily produced for its seeds, the seed yield is still important for the recognition and commercial viability of the cultivars on the market. Creating a cultivar of superior yield and forage quality with satisfactory seed production is one of the biggest challenges for alfalfa breeders and seed producers. The objective of this study was to determine forage and seed yields of 19 newly developed alfalfa experimental populations (ABP 1–19) of the Agricultural Institute Osijek during a long-term research period (2014–2018) in different climatic conditions. Significant differences were found between ABPs and years for forage and seed yields. Three-year (2014–2016) average green mass yield ranged from 68.41 t ha^–1 (ABP 6) to 78.05 t ha^–1 (ABP 19) and dry matter yield from 13.73 t ha^–1 (ABP 6) to 15.30 t ha^–1 (ABP 18). The average two-year (2017–2018) seed yield varied from 150.78 kg ha⁻¹ (ABP 9) to 335.35 kg ha⁻¹ (ABP 7). Annual forage yield significantly increased from the year of establishment to the second and third growing seasons of alfalfa. The highest average annual yield of green mass (90.24 t ha⁻¹) was achieved in 2015, dry matter yield (17.62 t ha⁻¹) in 2016 and the seed yield (394.17 kg ha⁻¹) in 2017. During the researched period there was a considerable decreasing trend in forage yield from the first to the last cut, except in the year of the alfalfa establishment. Several alfalfa populations (ABP 19, 8, 14) superior in all analyzed traits were identified, and they represent top performing materials with the potential for developing and releasing cultivars in the near future. Populations with high yields of green mass and dry matter (ABP 12, 18, 1) and seed yield (ABP 7, 4) were also detected and represent valuable genetic material to improve our alfalfa breeding program.

1. Introduction

Alfalfa (Medicago sativa L.) is one of the most important perennial forage crops for livestock production systems because of its high biomass yield and good nutritional quality, as well as desirable agricultural traits [1,2,3,4]. In addition to agronomic advantages, alfalfa also has many positive impacts on the environment in terms of enhancing soil fertility, protection against soil erosion, low N fertilization requirements due to its ability of biological nitrogen fixation and ability to provide nitrogen for the subsequent grain crop, reduction of energy and greenhouse gas emissions, and preservation of plant and animal biodiversity [5,6,7,8,9].

In Croatia, alfalfa is the most widely grown forage legume, with an average cultivated area for the 2013–2017 period of 23 162 ha and average annual hay production of 6.82 t ha^–1 [10]. In our agro-climatic conditions, alfalfa is grown in pure stands and usually persists over a 4–6 year period (with four to six cuts per year at 30–40 day intervals between cuts), depending on management practices, soil fertility, pests, weeds and weather conditions.

Forage yield with high nutritive value has been and continues to be a main objective of forage breeders [11]. Forage yield is a quantitative genetic trait controlled by genetic and environmental factors [12]. In general, genetic gains from breeding for yield in forages have been less than those achieved for grain yields in cereals [13]. Breeding progress for alfalfa yielding ability is hindered by several factors. These include the perennial nature of the crop (long selection cycles), small breeding investment, the harvesting of the entire plant (and hence the inability to make gains in harvest index), multiple harvests per year, as well as other factors such as high genotype by environment interaction, impossibility to select real hybrids or pure lines (owing to severe inbreeding depression), the high costs of phenotyping, tetrasomic inheritance, and the high level of non-additive variance [5,14,15,16,17]. Ren et al. [12] emphasize that in Canada the public and private investments in breeding of alfalfa are low compared to the other crops, and alfalfa breeding is dependent largely upon a few public breeding programs. The situation is similar in Croatia and most EU countries, and not only for alfalfa but also for other perennial forages.

Although alfalfa is not primarily bred for its seed, seed production is a prerequisite for good forage production as well as commercial exploitation of the crop. Seed yield is an important character for the market success of a new forage legume cultivar [18,19,20]. Creating cultivars of superior yield and forage quality with satisfactory seed production is one of the biggest challenges for alfalfa breeders and seed producers.

The objective of this study was to determine forage and seed yields of 19 newly developed alfalfa populations during a long-term research period in different climatic conditions, and to identify top performing materials, compared to the standard Croatian cultivar, as a favorable source for improvement of our breeding program and/or application of a potential future commercial cultivar.

2. Materials and Methods

2.1. Plant Material and Experimental Set-Up

Nineteen alfalfa experimental populations (ABP 1–19), which originated from different breeding cycles and applied methods, were evaluated during five years (2014–2018) of field research. The standard Croatian cultivar OS 66 was used as the control in the trial. All tested alfalfa materials were created within the framework of the alfalfa breeding program at the Department of Forage Crops Breeding and Genetics at the Agricultural Institute Osijek in Croatia.

The field experiment was established in the spring of 2014 at the Agricultural Institute Osijek using a randomized block (RCB) design with four replicates. All standard cultivation practices were applied (Figure 1a). The soil type at the growing site is eutric cambisol with 1.8–2% humus and a slightly acidic to neutral pH reaction (pH in KCl 6.4–7.0 with over 30 mg 100 g^–1 of soil P₂O₅ and K₂O), and is mostly used for setting up field trials in rotation with various crops. The plots were 1.2 m wide and 6 m long, with 0.2 m spacing between the rows. The sowing was performed by hand to a depth of 1.5 cm, with a planting rate of 15 kg ha^–1 (recommended rate for field scale production). No irrigation treatment was applied in the experiment. Weeds and pests were controlled using recommended pesticides when necessary.

2.2. Determination of Forage Yield

Populations/cultivar were evaluated through 15 cuts in total: four cuts in 2014, six cuts in 2015, five cuts in 2016, i.e., in the first to third year of the plant’s life. Green forage yield was measured by cutting the whole plot area to a 5 cm stubble height using a forage plot harvester with an electronic weigh system (Hege Model 212, Wintersteiger AG, Waldenburg, Germany, Figure 1b). Plants were harvested at the moment of late budding or at the stage of beginning of flowering. Immediately before cutting, subsamples of approximately 500 g of green mass were taken from the middle of each plot, weighed fresh, dried in a dryer at 105 °C for 48 h and weighed dry to determine dry matter content in order to calculate dry matter yield. Obtained data per plots were converted into tons per hectare (t ha^–1). Annual forage yields were determined by summing the cuts per years.

2.3. Determination of Seed Yield

The seed yield was determined from the second growth in the fourth and fifth productive year (2017, 2018) of alfalfa on all plots of the studied populations/cultivar. In both research years, the first cut was performed in the first half of May in order to make the alfalfa flowering coincide with the period of maximum abundance and activity of the main insect species pollinators (end of June–beginning of July), because alfalfa is primarily self-incompatible and pollinating insects are needed for cross-pollination among plants.

Alfalfa is an indeterminate flowering plant and in both years the flowering lasted from 20–25 June to the end of July (30–35 days), with the highest appearance of flowers in mid-July (Figure 1c). Harvesting of all plots was performed on the 20th of August 2017 and 2018 in the phase when most of the pods (80%–90%) have matured, i.e., when the pods on branches have acquired a dark brown color (Figure 1d). Desiccants were not applied. The plots were harvested with a Wintersteiger small-plot combine harvester (Figure 1e). Harvested seeds from all field plots were air-dried, cleaned using laboratory seed processing equipment (Figure 1f), weighed on an electronic scale and converted to seed yield in kilogram per hectare (kg ha^–1).

2.4. Climatic Conditions during the Study Period

During the alfalfa growing season (March-October) in 2014, the total monthly precipitation in all months, except for March and August, was higher than the long-term average (LTA, Figure 2a). May of 2014 was extremely rainy (161.4 mm), when 56% more precipitation was recorded compared to the LTA (70.8 mm). In 2015, the distribution of precipitation during the alfalfa growing season was extremely uneven. In April, June and July there was significantly less precipitation (from 58% to 79%), while in May, August and October significantly more precipitation (from 37% to 58%) compared to the LTA was recorded. The amount and distribution of precipitation per month in the 2016 growing season were similar to the LTA with slight deviations, except for July (110.8 mm, 44% more precipitation compared to the LTA) which had significantly more precipitation compared to the LTA (61.6 mm). In 2017, in all months, except for March and September, a lower or similar amount of precipitation was recorded compared to the LTA (82.6 mm), with the largest precipitation deficit in June (45.4 mm, 45% less precipitation than the LTA). In 2018, during most of the alfalfa growing season (April, May, August, September, October), there was less precipitation compared to the LTA (from 38% in August to 79% in October). Additionally, in the same year, 53% more precipitation was recorded in July (131.6 mm) compared to the LTA (61.6 mm). The total amount of precipitation in the alfalfa growing season in 2014 was 650.6 mm, which is by 158.4 mm, or 24% more than the LTA of 492.2 mm (Figure 2b). The growing season of 2016 was somewhat wetter, while the 2015 growing season was at the LTA level. In 2017 and 2018 the amount of precipitation recorded was below both the previously mentioned years and the LTA.

Mean monthly air temperatures during the alfalfa growing season were higher in most cases in all of the observed research years compared to the LTA (Figure 2c). Mean monthly air temperature deviation in relation to the LTA ranged from + 0.1 (July 2014) to + 5.4 °C (April 2018). In 2016, the mean monthly air temperatures were similar or slightly lower compared to the LTA, and the deviation ranged from −0.1 (May) to +1.5 °C (April). In 2018, in all months, with the exception of March, the air temperature was higher than the LTA, and April, May, August and October were particularly warm/hot with air temperature higher by 3.1 to 5.4 °C. In all research years, the mean air temperature of the alfalfa growing season was higher than that of the LTA, from 0.8 °C in 2016 to 2.4 °C in 2018 (Figure 2d). From the meteorological data of this research, a trend of increasing air temperature is visible, which can probably be attributed to global warming as well as a consequence of increasingly present climate change.

2.5. Statistical Analysis

All collected forage and seed yield data were processed using a two-factorial analysis of variance (ANOVA) with population and year as factors using the STAR v. 2.0.1 software [21]. Fisher’s protected LSD test was used at the 0.05 probability level to identify significant differences between the mean values of populations and years. To determine the impact of cuts on the green mass yield the data were systematized and processed by years.

3. Results and Discussion

Analysis of variance demonstrated statistically significant differences between experimental populations and years for green mass and dry matter yields (

Table 1. Results of two-way ANOVA for analyzed traits of 20 alfalfa experimental populations/cultivar.

Source of Variation	Degree of Freedom †	Green Mass Yield		Dry Matter Yield		Seed Yield
		MS	F-Value	MS	F-Value	MS	F-Value
Repetition	3	191.63	0.12 ns	7.12	0.15 ns	2987.24	0.81 ns
Year (Y)	2 (1)	59,665.83	3.60 **	2128.49	6.24 **	3,144,545.81	0.00035 **
Error year	6 (3)	65.90		2.82		9397.61
Population (P)	19	82.00	2.82 **	2.32	0.0001 **	22,264.34	0.0228 *
Y × P	38 (19)	17.61	0.67 ns	0.72	0.61 ns	10,375.16	0.616
Error	171 (114)	20.07		0.78		11,890.12
Total	239 (159)	527.03		18.83		32,436.00

^† Degrees of freedom given in parentheses ( ) are for the seed yield trait; MS—Mean square; * Significant at p < 0.05; ** Significant at p < 0.01, ns—not significant.

and

Table 2. Green mass and dry matter yields (t ha^–1) of alfalfa experimental populations/cultivar during 2014–2016.

Experimental Population	Green Mass Yield (t ha–1)				Dry Matter Yield (t ha–1)
	2014	2015	2016	Mean P	2014	2015	2016	Mean P
ABP 1	41.76	94.16	91.07	75.66 abcd	8.94	18.02	18.73	15.23 ab
ABP 2	40.74	86.48	87.62	71.62 ef	8.46	16.75	17.62	14.28 de
ABP 3	38.61	90.54	86.00	71.72 ef	8.23	18.08	17.07	14.46 cd
ABP 4	41.29	84.52	88.56	71.46 ef	8.67	16.57	17.60	14.28 de
ABP 5	40.99	88.71	87.62	72.44 de	8.24	16.88	17.38	14.17 de
ABP 6	38.84	85.42	80.98	68.41 f	8.13	16.65	16.42	13.73 e
ABP 7	38.91	87.49	88.97	71.79 ef	7.96	16.96	17.85	14.26 de
ABP 8	48.30	95.02	90.09	77.80 a	9.93	18.09	17.80	15.27 ab
ABP 9	40.42	90.71	89.16	73.43 cde	8.17	16.60	17.34	14.03 de
ABP 10	42.74	90.30	88.22	73.75 bcde	8.76	17.22	17.39	14.46 cd
ABP 11	42.68	88.10	84.79	71.86 ef	8.55	16.76	17.38	14.23 de
ABP 12	44.71	94.19	92.93	77.27 ab	8.93	17.10	17.89	14.64 abcd
ABP 13	40.51	91.97	85.66	72.71 cde	8.06	18.43	17.21	14.57 bcd
ABP 14	42.90	94.87	90.80	76.19 abc	8.38	17.27	17.85	14.50 cd
ABP 15	39.46	85.79	89.56	71.60 ef	7.82	16.95	17.64	14.14 de
ABP 16	41.80	87.13	91.75	73.56 cde	8.24	16.96	18.00	14.40 de
ABP 17	40.29	90.19	88.11	72.87 cde	7.99	17.07	17.30	14.12 de
ABP 18	47.47	96.40	88.89	77.59 a	9.62	18.42	17.86	15.30 a
ABP 19	46.74	94.61	92.80	78.05 a	9.15	18.16	18.18	15.16 abc
OS 66	43.51	88.13	88.28	73.31 cde	8.47	17.32	17.90	14.56 bcd
Mean Y	42.13 b	90.24 a	88.59 a	73.65	8.54 b	17.31 a	17.62 a	14.49
LSD 0.05	Year (Y): 3.14 **, Population (P): 3.61 ** Y × P: not significant, ** p ≤ 0.01				Year (Y): 0.65 **, Population (P): 0.71 ** Y × P: not significant, ** p ≤ 0.01

Lowercase letters—values followed by the same letter are not significantly different.

). The obtained results clearly indicate that environmental factors and genetic background of cultivars had a strong influence on observed traits. The highest three-year average green mass yield (78.05 t ha^–1) was achieved with ABP 19. High three-year green mass yields were also recorded with ABP 8, 12 and 18. All of the abovementioned experimental populations had a significantly higher yield, from 5.51% to 6.07% higher, compared to the yield achieved with the OS 66 cultivar which is used as a standard in field experiments for testing the economic value of newly reported materials in the official process of recognizing alfalfa cultivars in the Republic of Croatia. ABP 6 had the lowest average three-year green mass yield, which was 12.35% and 6.68% lower compared to the most productive/superior ABP 19 and the standard cultivar OS 66, respectively. The highest average three-year dry matter yield (15.30 t ha^–1) was achieved with APB 18, while high yield values were also recorded with ABP 8, 1, and 19; these were also higher compared to the standard cultivar (Table 2). ABP 6 had the lowest dry matter yield (13.73 t ha^–1), which was 10.26% lower compared to the most productive ABP 18.

The green mass yields obtained in this study are similar with the results of Cacan et al. [22], who determined variation in green mass yields from 61.69 to 84.29 t ha^–1 between local genotypes and cultivars of alfalfa evaluated in agroecological conditions in Turkey. However, our findings for the green mass yield were higher than those observed in the study reported by Turan [23]. Milić et al. [24] reported dry matter yields from 15.3 to 18.4 t ha^–1 in a three-year trial of commercial varieties and experimental populations in the second, third and fourth year of alfalfa stand life. Similar to Milić et al. [24], Avci et al. [25] and Song et al. [26], who studied forage production potential and nutritional values of superior alfalfa lines and cultivars of different geographic origin, obtained higher average three-year dry matter yields (from 16.14 to 19.89 t ha^–1 and 12.47 to 28.87 t ha^–1) than yields achieved in this research. Besides cultivar type, this variation in forage yields is probably caused by differences in applied cutting management systems, agronomical practices, years of alfalfa stand life included in the study, agroecological growing conditions of the area where the research was conducted, and the plot size used in the calculation of the yield per hectare.

In the second and third growing year of alfalfa, significantly higher average annual yields of green mass and dry matter were achieved (over 50% higher yield) compared to the year of alfalfa establishment (Table 2). Cavero et al. [27] also reported that the maximum alfalfa forage yield was lower in the first year (17 t ha^–1) than in the two following years (20–22 t ha^–1). This result is expected considering that alfalfa is a perennial crop and the expression of agronomic properties, in addition to the genetic potential of the cultivar, is strongly influenced by the stand age of the crop. In our agrological conditions, alfalfa achieves its maximum productivity potential in the second and third year of cultivation, and usually forage yield declines with increasing stand age. This result is in accordance with numerous previous studies where the effect of stand age on yield, nutritive value and persistence of alfalfa were studied [28,29,30,31].

Mean yield of green mass of all alfalfa populations/cultivar was significantly different between cuttings in each researched year (

Table 3. Mean green mass yield of all alfalfa experimental populations/cultivar per cuttings during 2014–2016.

Year	Cut	Date of Cutting	Mean Green Mass Yield (t ha–1)	Contribution of Cut in Total Annual Yield (%)
2014	I	20 June	8.56 c
	II	21 July	15.49 a
	III	26 Aug.	13.20 b
	IV	13 Oct.	4.88 d
LSD 0.05 0.82 **
	Total annual yield		42.13
2015	I	8 May	26.11 a
	II	8 June	17.59 b
	III	7 July	16.65 c
	IV	12 Aug.	10.51 e
	V	14 Sept.	13.03 d
	VI	3 Nov.	6.35 f
LSD 0.05 0.85 **
	Total annual yield		90.24
2016	I	9 May	27.64 a
	II	9 June	23.46 b
	III	15 July	15.26 c
	IV	16 Aug.	12.89 d
	V	29 Sept.	9.34 e
LSD 0.05 1.11 **
	Total annual yield		88.59

** p ≤ 0.01, Lowercase letters—values followed by the same letter are not significantly different.

). The highest average yield of green mass in the first production year was achieved in the second cutting (15.49 t ha^–1), which contributed 36.77% to the total annual yield. In the second and third growing year the highest yield was obtained in the first cutting (26.11 and 27.64 t ha^–1) which represented 28.94% and 31.20% of the total annual production, respectively. Alfalfa green mass yield decreased from the first to the last cutting, except in the first year. The last cut had the lowest yield in each growing year (4.88, 6.35 and 9.34 t ha⁻¹), accounting for 11.59%, 7.04% and 10.55% in the total annual yield. Similar results were reported by Wang et al. [32], who recorded the greatest values of forage yield in the first harvest and then a gradual yield decrease from the third to the sixth harvest which ranged from 3.4 to 4.3 t ha^–1 (averaged across 50 cultivars) and represented 10.8% to 15.2% of the annual total forage production. Djaman et al. [33] evaluated different fall dormancy-rating alfalfa cultivars for their forage yield during multiple years and found a decreasing trend in forage yield from the first cut to the fourth cut in each growing season. Numerous other researchers have also noted dominance of the first cutting in the total yield which can represent up to 30–50% of the annual forage yield of alfalfa [29,34,35,36,37].

Alfalfa forage yield depends to a great extent on various climatic factors, with precipitation and temperature the most important contributors [38]. A sufficient amount of precipitation in the year of establishment of alfalfa (2014), when it is most sensitive to lack of moisture, provided favorable conditions for optimal seed germination and emergence, as well as the growth and development of young plants during the growing season. The occurrence of drier periods and higher air temperatures during the growing season in the later growing years (2016, 2017) did not display a significant negative affect on the forage yield because alfalfa has a strong and deep root system that can absorb water from deeper soil layers, hence the plants can overcome stressful environmental conditions for a much longer period without negative consequences compared with many other common agricultural crops (Figure 2).

Analysis of variance revealed a significant influence of populations and years on alfalfa seed yield, while the interaction of population × year was not determined (

Table 4. Seed yield (kg ha⁻¹) of alfalfa experimental populations/cultivar during 2017 and 2018.

Experimental Population	Seed Yield (kg ha–1)
	2017	2018	Mean P
ABP 1	331.29	128.25	229.77 abcdef
ABP 2	323.27	115.69	219.48 cdef
ABP 3	416.78	68.57	242.68 abcdef
ABP 4	518.52	126.32	322.42 abc
ABP 5	457.94	135.06	296.50 abcd
ABP 6	293.04	79.23	186.14 ef
ABP 7	510.46	160.24	335.35 a
ABP 8	491.79	127.38	309.58 abc
ABP 9	203.27	98.29	150.78 f
ABP 10	309.12	69.92	189.52 def
ABP 11	353.78	105.80	229.79 abcdef
ABP 12	361.54	86.21	223.88 bcdef
ABP 13	333.48	107.23	220.36 cdef
ABP 14	479.22	133.98	306.60 abc
ABP 15	356.08	105.99	231.04 abcdef
ABP 16	396.73	72.17	234.45 abcdef
ABP 17	435.22	164.67	299.95 abc
ABP 18	439.85	131.98	285.91 abcde
ABP 19	510.20	147.75	328.98 ab
OS 66	361.89	111.08	236.49 abcdef
Mean Y	394.17 a	113.79 b	253.98
LSD 0.05	Year (Y): 48.77 **, Population (P): 108.00 * Y × P: not significant, ** p ≤ 0.01, * p ≤ 0.05

Lowercase letters—values followed by the same letter are not significantly different.

). The highest average two-year seed yield (335.35 kg ha^–1) was achieved with ABP 7. This population achieved a 28.88% higher seed yield compared to the yield achieved with the standard cultivar OS 66, and even 55.03% higher compared to ABP 9, which had the lowest yield (150.78 kg ha^–1). High average two-year seed yields were also recorded with ABP 19, 4, 8 and 14. In addition to genetic variation between populations, differences in seed yield were probably significantly influenced by abiotic environmental (e.g., temperature and precipitation) and biotic (e.g., pollinators and pests) factors during seed development.

In 2017, the average annual seed yield of all experimental populations/cultivar was 394.17 kg ha^–1, which was significantly higher, by 71%, compared to the yield achieved in 2018 (113.79 kg ha^–1). Considering the entire alfalfa vegetation period, both research years had a similar total amount of precipitation, which was slightly less than the LTA (Figure 1b). However, for the successful production of alfalfa seeds, the distribution of precipitation and air temperature during the summer months (June–August) plays a key role, because in those months the phases of flowering, pollination, flower fertilization, formation of pods and seeds, and seed filling and ripening occur. A significantly lower amount of precipitation in July of 2017, by 51.36% compared to the same month in 2018 (Figure 1a), probably had a favorable effect on the occurrence of higher numbers and activity levels of pollinating insects (higher degree of flower fertilization) as well as weaker growth (development of new shoots) and alfalfa lodging. Numerous authors have reported on the influence and importance of weather conditions on variation in alfalfa seed yield as well as other similar forage crops [39,40,41,42,43,44,45,46].

4. Conclusions

The results showed statistically significant differences between alfalfa experimental populations and years for forage and seed yields. Annual forage yield significantly increased from the year of establishment to the second and third growing seasons of alfalfa. During the researched period there was a considerable decreasing trend in forage yield from the first to the last cut, except in the year of alfalfa establishment. Several alfalfa populations (ABP 19, 8, 14) superior in all analyzed traits were identified and they represent top performing materials with the potential to develop and release cultivars in the near future. Populations with high yields of green mass and dry matter (ABP 12, 18, 1) and seed yield (ABP 7, 4) were also detected and represent valuable genetic material to improve our alfalfa breeding program.