North American Bristlegrass Seed Yield Response ... - Semantic Scholar

RESEARCH

North American Bristlegrass Seed Yield Response to Nitrogen Fertilizer and Environment Jorge A. López-García, William R. Ocumpaugh, J. Alfonso Ortega-Santos, John Lloyd-Reilley, and James P. Muir*

ABSTRACT Information on field management practices for seed production of newly domesticated grasses native to the southern Great Plains of North America has not been well documented. This study was conducted to document seed yield responses of one accession of plains bristlegrass [Setaria vulpiseta (Lam.) Roem. & Schult.] and three accessions of streambed bristlegrass [Setaria leucopila (Scribn. & Merr.) K. Schum.] to 0, 50, 75, 100, and 140 kg N ha−1 yr−1 under irrigation at Stephenville and Beeville, TX, during 2005 and 2006. The soil at Stephenville was a Windthorst fine sandy loam, mixed, thermic Udic Paleustalfs and at Beeville a Parrita clayey, mixed, active, hyperthermic, shallow Petrocalcic Paleustolls. Seed yields across N levels at Stephenville in year of establishment ranged (p < 0.05) from 57 to 753 kg ha−1 yr−1. October seed yield of accession 648 increased (p < 0.05) 305% with 75 kg N ha−1 compared to the no-N treatment at Beeville. Seed production peaked at 328 kg ha−1 for accession 648 and 352 kg ha−1 for accession 715 in spring 2006 at Beeville. Inflorescence density was positively correlated to seed yield (from 81 to 93%). Optimum bristlegrass N fertilizer management varied with accession, location, and year.

J.A. López-García, Rangelands and Forage Resources, INIFAP, Pasteur No. 414, Primer Piso, Querétaro, Querétaro, México 76040; W.R. Ocumpaugh, Forage Physiology and Management, Texas A&M Univ. AgriLIFE Research, 21643 Tyee Rd., Oakland, OR 97462; J.A. OrtegaSantos, Range Management, Texas A&M Univ.-Kingsville, MSC 218 700 Univ. Blvd., Kingsville, TX 78363; J. Lloyd-Reilley, Kika de la Garza Plant Materials Center, USDA-NRCS, 3409 N FM 1355, Kingsville, TX 78363-2704; J.P. Muir, Forages, Texas A&M Univ. AgriLIFE Research at Stephenville, 1229 N U.S. Hwy. 281, Stephenville, TX 76401. Received 25 Jan. 2010. *Corresponding author ([email protected]). Abbreviations: PCA, principal component analysis.

G

rasses native to the southern Great Plains of North America are important not only for grazing cattle but also for restoring wildlife habitat for northern bobwhite quail (Colinus virginianus) and wild turkey (Meleagris gallopavo). Nature tourism and hunting has increasing relevance because ranchers rely on these for extra sources of income. In 2001, people participating in activities related to wildlife in Texas spent $5.2 billion in retail sales, generated $2.6 billion in salaries and wages, and contributed approximately 98,000 jobs to the state (Texas Department of Agriculture, 2007). Protecting soil from erosion is an additional vital role that native grasses play. Cited by Harper-Lore and Wilson (1999), the Presidential Executive Memorandum “Environmentally and Economically Beneficial Practices on Federal Landscaped Grounds”, issued on April 26 1994, established that all federally funded projects, including landscaping and road construction, are required to use regionally adapted native plants. The stated goal is to protect natural resources and reduce the use of agrochemicals. The same authors mentioned that planting native grasses and forbs not only Published in Crop Sci. 51:361–369 (2011). doi: 10.2135/cropsci2010.01.0036 Published online 15 Nov. 2010. © Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

CROP SCIENCE, VOL. 51, JANUARY– FEBRUARY 2011

361

Table 1. Soil physicochemical characteristics at the experiment locations in Texas. Stephenville Value Qualifier

Beeville Value Qualifier

pH† Electrical conductivity (dS m−1)

4.9 0.248

6.5 0.547

Nitrate-N (mg kg−1)‡ P (mg kg−1)§ K (mg kg−1) Texture¶ Sand (g kg−1) Silt (g kg−1) Clay (g kg−1) Textural class Organic matter (g kg−1)

65.0 37.0 149.0

Characteristic

High Moderate Moderate

650 130 220 Sandy clay loam 10.7

99.0 25.0 169.0

Very high Low High

710 70 220 Sandy clay loam 21.9

†

pH and electrical conductivity were determined in a 1:2 soil:water extract of soils. Nitrate-N was extracted from soils using a 1 N KCl solution. § P and K were extracted from soils using the Mehlich III extractant. ¶ Texture was determined by the hydrometer procedure. ‡

stabilizes soil and prevents erosion but also provides aesthetically pleasing plant cover along roadsides. Previous work in the southern Great Plains identified several plains bristlegrass accessions as valuable because of their survival and spring regrowth as well as their forage mass and seed production potential in several Texas environments ( J. Lloyd-Reilley and S.D. Maher, unpublished data, 2007). The screened bristlegrass accessions native to south Texas include streambed bristlegrass [Setaria leucopila (Scribn. & Merr.) K. Schum] and plains bristlegrass [Setaria vulpiseta (Lam.) Roem. & Schult.] (USDA-NRCS, 2007). Streambed bristlegrass is a perennial, cespitose plant with culms 20 to 100 cm in height. Its inflorescence is pale green, almost cylindrical, and has closely packed panicles 6 to 15 cm long, although they may be shorter. This grass occurs naturally in the drier regions of Texas but has been found in almost all of the southwestern United States and northern Mexico (Gould and Kapadia, 1975; Correl and Johnston, 1996; Barkworth et al., 2003). Streambed bristlegrass has high forage and seed production compared to other native grasses (Hatch et al., 1999). Plains bristlegrass is also a perennial warm-season grass that grows naturally in Texas, New Mexico, Arizona, Colorado, and Mississippi (USDA-NRCS, 2008a). It is an erect plant reaching 90 cm height and grows during the spring and summer. Plains bristlegrass is adapted to medium textured soils with pH between 6 and 8 but does not tolerate salinity. This plant is found in regions with an annual precipitation between 300 and 600 mm and produces highly palatable forage for grazing animals and bears seed in medium abundance (USDA-NRCS, 2008b). The importance of N fertilization for seed production in grasses has been well documented. Humphreys and Riveros (1986) stated that seed production of various tropical grasses increased when N fertilizer was applied, since this nutrient 362

plays a crucial role regulating the speed at which several mechanisms lead to seed development. They also added that the first step in a research program is to determine the grass seed production response to N fertilization. However, unimproved grasses tend to have less response to soil N than those selected for cultivation (Abraham et al., 2009); accessions of native or unselected species also show wide ranges of seed production response to soil fertility (Chivers and Aldous, 2005). Because of the significance of how new species and accessions within species respond to N, it is important to determine optimal N fertilizer levels for new crop seed production (Loch et al., 1999b). Gislum and Griffith (2004) stressed the importance of N fertilization in the development and acceleration of reproductive tiller emergence in grass plants. For example, the number of tillers per plant is augmented with increasing N fertilizer levels applied to perennial ryegrass (Lolium perenne L.). Agronomic information on seed production from native Setaria spp. in Texas is almost nonexistent and, as the demand for their seed by ranchers and public organizations in North America rises, it is essential to develop technology to effectively produce seed. The objective of this study was to document seed production response to N fertilizer levels of four bristlegrass accessions in two locations in Texas.

MATERIALS AND METHODS Experiments were established in Beeville (10 May 2005) and Stephenville (2–5 May 2005), TX, during spring 2005. The sites differed in temperature and soils (Tables 1 and 2). Stephenville is located at 32°15´ N and 98°11´ W with an altitude of 407 m, while Beeville is located at 28°27´ N and 97°42´ W with an altitude of 71 m. Soil taxonomic classification was obtained from the “Web Soil Survey” (USDA-NRCS, 2008c); soil physicochemical characteristics at the experimental locations appear in Table 1. The experimental design was a split plot with two factors at both Beeville and Stephenville with bristlegrass accessions (four) as whole-plot and N levels (five) as subplots. Plots were set up in the field in a randomized complete block design with two replications. Bristlegrass accessions 9029648, 9029677, 38715, and 9038820 were originally collected at Webb, Karnes, Duval, and Willacy Counties, TX, respectively. These accessions will be denoted subsequently as 648, 677, 715, and 820, respectively. Seeds of these bristlegrass accessions were germinated under greenhouse conditions in plastic trays with individual containers 9 cm tall, 7.5 cm deep, and 7.5 cm long fi lled with moss peat. Trays were allocated on benches 90 cm high. Accession 648 is a Setaria vulpiseta and the remainder are S. leucopila. Nitrogen fertilization treatments imposed in 2005 and 2006 consisted of 0, 50, 75, 100, and 140 kg ha−1 applied to the same plots each year. Urea was used as the source of N and it was evenly broadcasted by hand onto each plot. The N levels were split in two in the 2006 growing season to span the bristlegrass growing season; the fi rst half was applied 6 Apr. 2006 and the second half on 4 Sept. 2006. Along with the N treatments, both experimental sites received 60 kg ha−1 of P and 60 kg ha−1 of K in 2005 and 2006 to ensure that these macronutrients were not limiting. Each plot consisted of three rows of plants spaced 30.5 cm apart. Interplant spacing was 30.5 cm, resulting in 20 plants

WWW.CROPS.ORG


Table 2. High, low, and average air temperatures and rainfall at Stephenville and Beeville, TX, during the study. Stephenville

Month January February March April May June July August September October November December Average †

2005 Temperature (°C) High Low Avg 15 15 19 24 27 33 34 34 34 26 23 15 25

2 3 5 9 16 20 21 21 18 10 5 −2 11

8 9 16 17 21 27 28 27 26 18 14 7 18

Rain mm 45 72 60 1 88 11 60 55 31 29 0 2 456†

Beeville

2006 Temperature (°C) High Low Avg 20 15 22 28 30 34 36 37 30 26 21 15

2 1 9 13 16 19 22 23 16 9 7 1

11 8 16 21 23 26 29 30 23 18 14 8


8 11 12 15 18 21 23 23 22 15 12 6 15

14 15 17 21 24 27 29 29 29 22 19 13 22

Rain mm 102 89 192 25 47 88 83 47 4 24 8 2 696†


8 6 14 18 21 22 23 23 21 16 11 8 16

16 14 21 24 28 28 28 30 26 23 19 14 22

Rain mm 12 11 8 0 104 84 117 22 226 43 1 63 690†

Rain is expressed in terms of annual total.

within each 6.1-m row in 5.6 m 2 plots. Only the center row was used for data collection. Plots were placed 1 m apart. The experiments were sprinkler irrigated whenever plants wilted, although this was changed to drip irrigation at Beeville in 2006 to avoid impacting seed shattering differentially, since accessions matured unevenly. Seed harvest, initiated at individual accession seed maturity peak, occurred from 21 to 24 November at Beeville and 29 July to 3 August at Stephenville in 2005; 2006 harvests took place from 4 May to 2 June and from 28 September to 5 October at Beeville while no harvest took place at Stephenville because of poor regrowth after the 2005–2006 winter season low temperatures. Plants were clipped to 30-cm stubble height after every seed yield sampling to make regrowth uniform within plots and simulate mechanical seed harvest. Harvest consisted of clipping reproductive stems with scissors and placing them in poly-woven bags to allow air circulation. As all the Setaria accessions included in the study exhibited seed shattering traits, maturity was closely monitored in each plot to harvest mature seed and minimize seed loss. Floral stems from three plants in every plot were set apart in paper bags to quantify inflorescence density and length. Bags containing the samples were placed in forced-air driers with no added heat until the plant material was dry. Seed was shattered from panicles by hand. Structures containing a fully developed caryopsis were separated from those having no or immature caryopsis using an air column blower. Each sample was blown for 2.5 min at an aperture of 3 cm at the column top because this was determined to be optimal in previous tests. The proportion of fi lled seed, defi ned as structures containing well-developed caryopses, was calculated as the percentage of fully developed caryopses contained in the total mass harvested (fully developed caryopses plus empty structures). Average inflorescence length was calculated from 10 randomly selected panicles per plot. This sample size was determined by the Stein’s equation (Steel and Torrie, 1980) with a presample size of 103 inflorescences. To reduce insect and ergot damage, Malathion [diethyl (dimethoxy thiophosphorylthio) succinate] 56% a.i. (at a rate of 1.75 L ha−1) and Tilt (propiconazole) (Syngenta Crop Protection CROP SCIENCE, VOL. 51, JANUARY– FEBRUARY 2011

Inc., Greensboro, NC) 41.8% a.i. (at a rate of 293 mL ha−1) were applied to the field plots when insects or ergot were identified. Maximum and minimum daily temperatures at both experimental sites were obtained from the NOAA Satellite and Information Service web page (National Climatic Data Center, 2008) for the Texas Beeville (5 NE 1894-06-200801 [410639]) and Stephenville (1N 1918-01-2008-01 [418623]) Texas AgriLife centers (Table 2). Beeville daily rainfall data was taken from the former source, and information for Stephenville was acquired from the Texas AgriLife Research web page for the Texas AgriLife Research and Extension Center at Stephenville (R. Wolfe and J. Brady, unpublished data, 2008). Daylength at both sites was computed from sunrise and sunset tables provided by U.S. Naval Observatory (2008). Results from each environment were analyzed separately due to distinct conditions at each location and year. Data were analyzed using PROC MIXED of SAS (SAS Institute, 2004). Each site year was considered an independent environment due to differences in edapho-climatic conditions. Replication and replication × accession were included as random error terms in the RANDOM statement in the split plot design analysis (Littell et al., 2002). Least square means of the treatments were calculated and differences between them were estimated by the PDIFF option of the LSMEANS statement as proposed by Littell et al. (2002) for a MIXED procedure. Differences between means were considered significant at the p ≤ 0.05. Arcsin square root transformation of proportion of fi lled seed data was performed before ANOVA and mean comparisons were done. The original proportion of fi lled seed data are shown in results. Principal component analysis (PCA) has been used to detect correlations between crop agronomic characteristics and performance. For instance, PCA was performed to establish a relationship between yield and other variables when N fertilizer treatments were studied in forage production ( Jefferson et al., 2001). In our study, PCA was performed to discriminate any relationships among the following dependent variables: seed yield (kg ha−1), inflorescence density, inflorescence length (cm), and proportion of fi lled seed. Criteria for selecting

WWW.CROPS.ORG

363

Principal component analysis was performed by the SAS PROC FACTOR statement, including the METHOD = PRIN and ROTATE = VARIMAX options (O’Rourke et al., 2005).

RESULTS Seed Yield Stephenville At Stephenville, it was possible to report seed yield only once in 2005 because low temperatures during the 2005– 2006 winter damaged grass plants at the experimental site, and 42% of the plots did not regrow. Even at the end of March 2006, temperatures between −2.8 and 1.7°C were recorded during five continuous days (Table 2). There was an interaction between accessions and N levels. Seed yield of accession 820 increased with N fertilizer application at 75 kg ha−1 or more. In this accession, seed yield was 210% greater on average for 140 kg N ha−1 compared to the control. Type of response of this accession to N fertilizer levels is presented in Fig. 1. No differences were detected in seed production of this accession between the control and 50 kg N ha−1 treatments (Table 3). In accession 715 seed yield was 170% greater when N fertilizer was applied at 50 or 75 kg ha−1 compared with the control but yields decreased at the higher fertilizer levels so that they were no different from the control. Accessions 648 and 677 seed yields did not respond to N fertilizer. Seed yield in July 2006 in surviving plots averaged 23 kg ha−1.

Figure 1. Response of bristlegrass 820 accession seed yield to N fertilization in Stephenville, TX, (a) in July 2005 and Beeville, TX, (b) in October 2006.

components to be retained for interpretation were eigen values of components ≥1 (with the exception of component 2 of data from July 2005 at Stephenville with an eigen value of 0.7) and a combined cumulative percent of variance ≥70%. The principal axis method was applied to extract the components, followed by a variomax rotation. A dependent variable was included in a component if the loading was ≥0.60 for that component. Table 3. Seed yield of bristlegrass accessions harvested in July 2005 as a response to five nitrogen fertilizer levels at Stephenville, TX. N rate

648

Accession 677 715

820

––––––––––––––––––––––––––––––kg ha−1–––––––––––––––––––––––––––––– 132a 116b 243c 0 103a† 50 128a 65a 335a 359c 75 57a 147a 290a 562b 100 88a 123a 278ab 592ab 140 102a 115a 264ab 753a SE 58 58 58 58 †

Beeville Beeville seed harvest occurred in November 2005 and May and October 2006. In November 2005, average seed production differed among accessions. Accession 715 produced 168 kg ha−1, accession 648 produced greater than 93 kg ha−1, accession 677 produced 121 kg ha−1, and accession 820 produced 67 kg ha−1. The addition of N had no effect on any accession. In May 2006, accessions 648 (328 kg ha−1) and 715 (352 kg ha−1) reached their greatest seed yield. Accession 820 yielded 143 kg ha−1, the lowest seed production, and accession 677 had an intermediate seed yield of 231 kg ha−1. Nitrogen fertilization did not affect seed yields at this harvest. In October 2006, an accession × N rate interaction occurred (Table 4). The most extreme effect of the interaction occurred with accession 648 for which 75 kg N ha−1 increased seed yields 305% compared with the control; however, greater N levels diminished seed yield compared to the 75 kg N treatment. By contrast, seed yield of accessions 677 and 715 increased (307 and 95%, respectively) with application of 140 kg N ha−1 compared with the control. Conversely, accession 820 did not respond to increasing N levels (Fig. 1). An interaction of accession × harvest date was detected when seed yield of the three harvest dates at Beeville where compared. Seed yield of accessions was greater in May 2006

Means in each column followed by the same letter are not different at p < 0.05.

364

WWW.CROPS.ORG


than November 2005 and October 2006 and no differences were detected between the last two harvest dates (Table 5). Seed maturity and harvest occurred on a continuum in the summer 2006 at Beeville. Seed was harvested between 4 and 9 May in accession 648, between 9 and 15 May in accession 677, between 13 and 16 May in accession 715, and between 29 May and 2 June in accession 820. Seed of accessions matured in a shorter range at Stephenville. The second harvest seed maturity at Beeville fell completely within November in 2005 and October in 2006.

Table 4. Seed yield of bristlegrass accessions harvested in October 2006 as a response to five N fertilizer levels at Beeville, TX. N rate

Principal components are displayed as columns under the heading “Component” in Tables 7 and 8. A variable with a ≥0.60 value was considered as having a positive or negative relationship to that component, and that value represents the loading a given variable had on that component. In all cases, component 1 (first column) consists of variables related to seed production; component 2 (second column) includes variables not related to seed production. Based on PCA using data from both Stephenville and Beeville, seed yield and inflorescence density were positively related to seed production while inflorescence length and proportion of filled seed were mostly inversely or unrelated to seed production. Stephenville The two principal components retained in the PCA for this site together explained 82% of the total variance. Seed yield and inflorescence density displayed a high relationship to component 1, whereas inflorescence length was not related to seed production as it had a high loading to component 2 (Table 7). The positive correlation of seed yield and inflorescence density with seed production is explained primarily by the interaction between accessions and N levels. The positive loading of seed fi ll proportion with seed production at Stephenville (Table 7) is explained by the difference between the percent of seed fi lled by accessions. Accessions 820 and 715 had a greater proportion of fi lled seed compared to accessions 677 and 648 (Table 6). Likewise, all accessions followed the same pattern for seed

Accession 677 715

820

––––––––––––––––––––––––––––––kg ha−1–––––––––––––––––––––––––––– 45c 88b 41a 0 60b† 50 76b 81bc 88b 70a 75 243a 100b 98b 72a 100 87b 75bc 128ab 72a 140 107b 183a 172a 55a SE 17 17 17 17 †

Relationship between Seed Yield and Seed Yield Components

648

Means in each column followed by the same letter are not different at p < 0.05.

yield. Accession 820 yielded the most, followed by 715, while 677 and 648 were the least productive. Inflorescence length was not related to seed production (Table 7), which is explained by the dissimilarity between the seed yield and the inflorescence length among accessions. Although accessions 648 and 677, with an average length of 23 cm, had longer panicles than the other accessions, both accessions yielded the lowest quantity of seed. By contrast, accession 820, which produced the most seed, had inflorescences that averaged 17 cm while accession 715, with medium seed yield, had the shortest inflorescences (13 cm) (Table 6). Beeville November 2005. The two principal components rotated from the data explained 79% of the total variance (Table 7). Seed yield and inflorescence density were positively correlated with seed production (component 1), but proportion of fi lled seed showed a negative relationship to this component; by contrast, inflorescence length had a high loading to component 2, so was not related to seed production. The difference in inflorescence density among accessions accounted for the relationship between this variable and seed production. Inflorescence density was similar to seed yields for most accessions. Accession 715, which produced the greatest quantity of seed, had more panicles than the other accessions. Accession 820, which yielded less seed than the other accessions, also produced fewer inflorescences (Table 6).

Table 5. Seed yield comparison of bristlegrass accessions harvested at different dates at Beeville, TX.

Accession 648 677 715 820 SE

November 2005

Seed yield May 2006

October 2006

Level of significance November 2005 vs. May 2006 vs. November 2005 vs. May 2006 October 2006 October 2006

––––––––––––––––––––––kg ha−1–––––––––––––––––––––– 93 328 114 121 231 97 168 352 115 67 143 62 24 24 24

*** ** *** *

*** ** *** *

NS† NS NS NS

*Significant at the 0.05 probability level. **Significant at the 0.01 probability level. ***Significant at the 0.001 probability level. † NS, not significant. CROP SCIENCE, VOL. 51, JANUARY– FEBRUARY 2011

WWW.CROPS.ORG

365

Table 6. Seed yield components of bristlegrass accessions for several harvest dates at Stephenville and Beeville, TX. 648

Variable Inflorescence density (in panicles m−2) Filled seed (%) Inflorescence length (in cm) Inflorescence density (in panicles m−2) Filled seed (%) Inflorescence length (in cm) Inflorescence density (in panicles m−2) Filled seed (%) Inflorescence length (in cm) Inflorescence density (in panicles m−2) Filled seed (%) Inflorescence length (in cm) †

Accession 677 715 820

SE

Stephenville July 2005 192b† 160b 295a 299a 9 37d 48c 62b 72a 22 21b 25a 13d 17c 1 Beeville November 2005 853b 687bc 1147a 521c 34 56c 66b 57c 77a 17 10b 11a 6d 8c 0.2 Beeville May 2006 430b 327b 618a 194c 8 78a 79a 75a 46b 17 10c 12b 8d 15a 0.2 Beeville October 2006 142b 123b 212a 110b 4 72 77 68 59 20 9ab 10a 7c 8b 0.3

Means in each row followed by the same letter are not different at p < 0.05.

Proportion of fi lled seed differences among accessions helps explain the negative loading of this variable to component 1 (Table 7). Proportion of fi lled seed in accessions 715 and 820 was unrelated to the pattern of their seed yield. For example, accession 820 fi lled-seed percentage was greater than that of the other accessions while the percent seed fi lled in accessions 648 and 715 was the lowest. However, accessions 715 and 820 produced the greatest and the least amount of seed, respectively, while accession 677 fi lled-seed percentage was intermediate as was its seed production yield (Table 6). Accession inflorescence length ranged from 6 to 11 cm. For most accessions, inflorescence length was not correlated to seed production so it had a high loading to component 2 (Table 8). Accessions 648 and 677 bore longer panicles than the other accessions yet they differed from each and their seed yield was intermediate. Accession 715, with the shortest spikes, attained the greatest seed production. Accession 820 had intermediate inflorescence length but produced the lowest seed yield (Table 6).

May 2006. The total variance explained by the data set

for the two rotated components was 82%. Seed yield and inflorescence density from plants were positively correlated to seed production. In contrast, inflorescence length of accessions had a negative relationship to seed production. Proportion of fi lled seed was positively associated to component 2 and thus was not related to seed production (Table 8). The positive correlation seed yield and inflorescence density had with seed production came from differences among accessions (Table 8). Accession 715 produced more inflorescences than the other accessions and there were no differences between accessions 648 and 677; however, both had greater numbers of inflorescences than accession 820. Furthermore, accession inflorescence numbers had a similar response to that for seed yield, although seed production of accession 715 was not greater than that of accessions 648 and 677 (Table 6). The negative loading inflorescence length had with component 1 (seed production) is explained by the dissimilarity shown by accession panicle length and its relationship to seed yield (Table 8). Accession 820 had the lowest seed yield and the longest inflorescences compared with the other accessions. Similarly, accession 677 had longer panicles than accessions 648 and 715, but its seed yield was lower than these accessions. By contrast, accession 648 yielded the most seed but produced midsize inflorescences. Accession 715 ranked second in seed production but it developed the shortest inflorescences compared with the other accessions (Table 6). Proportion of fi lled seed was not related to seed production as this variable loaded to component 2 (Table 8). Accessions 648, 677, and 715 fi lled between 75 and 79% of their seeds and there were no differences among them; accession 820 had 46% of fi lled seed. This last accession attained the lowest seed yield, which in turn was not different from accession 677, which had a high proportion of fi lled seed (Table 6). October 2006. The cumulative variance explained by

the two components rotated for the data set was 73%. Table 7. Rotated factor pattern and final communality estimates of component 1† (C1) and component 2 (C2) extracted from principal component analysis between variables under study at Stephenville and Beeville in 2005.

Variable

Stephenville July 2005 C1 C2 h2‡

Seed yield 0.93 Inflorescence density 0.81 Inflorescence length −0.18 Proportion filled seed 0.71 †

−0.20 −0.12 0.96 −0.48

0.90 0.67 0.95 0.73

Beeville November 2005 C1 C2 h2 0.81 0.91 −0.22 −0.69

−0.31 −0.14 0.86 −0.54

0.75 0.85 0.79 0.76

Component 1 includes variables related to seed yield and component 2 includes variables not related to seed yield. ‡ Communality estimates are denoted by h2.

366

Table 8. Rotated factor pattern and final communality estimates of component 1† (C1) and component 2 (C2) extracted from principal component analysis between variables under study at Beeville in 2006. Beeville May 2006 C1 C2 h2‡

Variable Seed yield Inflorescence density Inflorescence length Proportion filled seed

0.86 0.23 0.88 0.26 −0.73 −0.56 0.26 0.90

0.79 0.85 0.84 0.88

Beeville October 2006 C1 C2 h2 0.89 0.93 −0.12 0.15

0.35 −0.23 0.80 0.66

0.90 0.92 0.65 0.45

†

Component 1 includes variables related to seed yield and component 2 includes variables not related to seed yield. ‡ Communality estimates are denoted by h2.

WWW.CROPS.ORG


Seed yield and inflorescence density were correlated positively to seed production and settled in component 1; inflorescence length and proportion of fi lled seed loaded onto component 2, so they were not related to seed production (Table 8). The panicle density of accession 715 was greater than the other accessions (Table 6). Also, more panicles were produced with 75, 100, and 140 kg N ha−1 than with the control, but yields at the 50 kg N ha−1 rate were similar to the control and other N levels. In general, accessions that formed more inflorescences produced more seed, and the same response occurred for N levels and seed production (Table 4), which explains why the density of inflorescences loaded onto component 1 (Table 8). Inflorescence length of accessions did not correlate to their seed yield (Table 4). Thus, accession 715 had the shortest inflorescences compared with the others but reached the greatest seed yield on average. By contrast, accession 648 formed longer inflorescences than accession 715 but had similar seed yield. Accession 677 attained, on average, a lower seed yield but bore longer panicles than accession 715. Accession 820 formed panicles that were intermediate to the other entries but had the lowest seed yield (Table 6). When response of inflorescence length to N fertilization levels was compared, 140 kg N ha−1 produced longer panicles than the other N levels (9.7 vs. 8.5 cm, on average); however, average seed yield of 0 and 50 kg N ha−1 was lower than the other N levels. This is the reason why the inflorescence length variable was not related to seed production and loaded onto component 2 (Table 8). The proportion of fi lled seed from plants varied between 59 and 77% and no differences were detected among them (Table 6). However, seed yield was different between accessions. Therefore, similarity in proportion of fi lled seed among accessions explains why this variable was not related to seed production (Table 8).

DISCUSSION Seed Yield The effect of fertilizer N levels on bristlegrass seed yield differed among accessions at the two experiment sites likely due to differences in growing environment. Variation was also observed between years and seasons at Beeville. This type of variation has been reported before when cultivars of Setaria spp. were evaluated at different seasons and at several N levels ( Jank and Hacker, 2004) as well as within accessions of weeping grass [Microlaena stipoides (Labill.) R. Br.] (Chivers and Aldous, 2005). Humphreys and Riveros (1986), discussing other grass species, reported that as N fertilizer levels increased seed yield also increased until a maximum response was reached after which further N levels diminished seed production. Jank and Hacker (2004) also stated that N fertilization commonly had a positive effect on the seed yield CROP SCIENCE, VOL. 51, JANUARY– FEBRUARY 2011

of different Setaria selections but did not report a decline at excessive N levels. Moore et al. (2004) emphasized the important role daylength plays in the physiology and development of warm season grasses but in floral induction in particular. Loch et al. (1999a) analyzed information about photoperiod requirements for flowering of several grasses studied under controlled conditions, in which Setaria sphacelata (Schumach.) Stapf & C. E. Hubb. cv. Nandi was classified as a quantitative long-day species. Loch (1980) referred to quantitative response to daylength as “critical daylength promotes, but is not essential for flowering.” Furthermore, Humphreys and Riveros (1986) affirmed that this Setaria grass will flower in a wide range of photoperiods (from 8 to 24 h), but they added that flowering will be enhanced under the longest daylengths; however, seed production of quantitative daylength plants may be diminished if daylength is not suitable (Fisher, 1999). Seed production of accessions in the study was greater in Stephenville than in Beeville in November 2005 as well as May 2006. This response may be due to differences in temperature, daylength, and rainfall. Temperature did not appear to limit plant growth or seed production, as the mean temperature was 27°C at Stephenville and 24°C at Beeville 2 mo before harvesting. However, there were 10 d with temperatures between 15.5 and 10.5°C at Beeville the month before harvesting. This may indicate that temperatures at Stephenville were more appropriate for seed production for the species evaluated on this study. Lower mean temperatures may have diminished seed production in October and November compared with seed yields attained in May at Beeville, as lower maximum and minimum mean temperatures were observed in the previous months (Table 2). Low temperatures at night can decrease inflorescence production, diminish floret growth and maturation, and negatively affect seed set (Loch, 1980). Furthermore, variation of environmental conditions can play a major effect on the variability of seed yield among years, even greater than soil variability or management practices applied to several crops (Eghball and Varvel, 1997; Lamb et al., 1997). Differences in seed production between the two experiment sites in 2005 may have arisen from a longer daylength at Stephenville than at Beeville. Days were longer by 2 h (14 vs. 12 h) 2 mo before harvest to 3 h (14 vs. 11 h) 1 mo before harvest at Stephenville than in Beeville. This difference of daylength between the two sites was 3 h (13 h 22 min vs. 10 h 27 min) at the time of harvest. A similar situation may have occurred with inflorescence density and seed yield in 2006 at Beeville. Harvest time in May occurred throughout the entire month (from the 4 May to the 2 June) for most of the plots. Grass plants were harvested during a shorter period in October from 28 September to 5 October. For that reason, daylength is reported as a period for May and as a single date for

WWW.CROPS.ORG

367

October. Plants received 30 min to 1 h more daylight 1 mo before they were harvested in May (from 12 h 25 min to 13 h 06 min) than those harvested in October (11 h 58 min). Similarly, daylength was 2 to 2.5 h longer during the period of harvesting in May (from 13 h 12 min to 13 h 40 min) than in October (11h 8 min). However, 2 mo before harvest, daylength was 32 min to 1 h 17 min longer for plants harvested in October than those harvested in May. Even though the experiments were conducted under irrigated conditions, rain distribution appeared to have played a role in the difference in seed production among the two sites. One hundred and eighty-seven millimeters of rain were well distributed at Stephenville during seed fi ll and maturation. Conversely, 112 mm were recorded at Beeville around the same seed developmental stages, but 83% of that fell on 1 d at the end of the harvest season. A similar positive correlation between natural occurrence of Panicoideae subfamily members and annual precipitation (0.89) has been observed in North America (Taub, 2000). Although seed production in May 2006 was greater than in October 2006 at Beeville (Table 5), positive response to N fertilization compared with the control, where it occurred, was superior later in the year and increased from 5 to 144% in May to 75 to 307% in October.

Relationship between Seed Yield and its Components Inflorescence Density Inflorescence density was positively correlated with seed yield. Their relationship to accessions, which was identified as related to seed production, ranged between 81 and 93% throughout the locations and harvest dates in this study. The average panicle density varied between 147 and 792 per m−2. This last density was produced in November 2005 in Beeville, and it may reflect the high levels of available N status at the beginning of the study (Table 1). The average inflorescence density per m−2 at Beeville decreased to 413 in May 2006 and then to 140 in October 2006. A better production of panicles in May compared to the October production can be attributed to differences in the mean temperature during the previous and current month of harvest. The mean low temperature in May was 19.5°C, which was 6°C higher than the October mean low temperature and hence with a better benefit for inflorescence formation. A positive relationship between the seed yield and the panicles density was reported by Humphreys and Riveros (1986). These authors also mentioned that this phenomenon is common in tropical grasses. Another factor that may have affected the panicle production between places and dates was daylength. This important environmental factor may have influenced panicle production the same way it did seed yield, and has been discussed previously. Although the inflorescence density in July 2005 at Stephenville was lower than that obtained at Beeville in 368

May 2006, the average seed production across the accessions was greater in the fi rst location. This was apparently due to the contribution of seed yield components other than the ones measured in this study. Inflorescence Length Accession inflorescence length was unrelated or had a negative relationship to seed production. This correlation was mainly a result of the differences in the inflorescence length of the accessions in the study and the fact that accessions bearing longer inflorescences also attained less seed yield. Although 140 kg N ha−1 produced longer panicles than the other N levels at Beeville in October 2006, in general, N levels did not affect this variable across sites and dates of evaluation. Loch et al. (1999c) reported that panicle size of some grasses can be increased by N fertilization but this response is not as common as increase of inflorescence density. It is possible that inflorescence length is controlled more by genetic factors and is less impacted by management factors. Proportion of Filled Seed A positive (71%) correlation between seed production and the proportion of fi lled seed of the accessions was registered at Stephenville. By contrast, seed production had a negative (−69%) relationship with the proportion of fi lled seed at Beeville in November 2005. However, this variable was not related to seed yield at Beeville in May and October 2006. This phenomenon was also observed by Young et al. (1999), who reported that N addition had a positive effect on orchardgrasss (Dactylis glomerata L.) seed yield and total dry weight of harvested reproductive structures (registered before threshing); however, tall fescue (Festuca arundinacea Schreb.) seed production was unaffected while increasing total dry weight of those structures, indicating that seed yield is not necessarily related to the total dry weight of the harvested material and that other seed yield components may be responsible for seed yield changes.

CONCLUSIONS Some accessions produced greater yields at the northern site while others were more productive in the south. Soil N fertilization should therefore be applied according to individual accession requirements and site differences including soil fertility and texture, temperature, daylength, and soil moisture. Among the seed yield components studied, inflorescence density exerted the greatest influence on seed yield and can be used in plant or population selection for greater seed yield. Seed yield response to N fertilization varied considerably among accessions: low levels increased yields in some whereas high levels were more effective for others while a few had no response at all. Soil N fertilizer levels should therefore be tailored to each accession.

WWW.CROPS.ORG


References Abraham, J.K., J.D. Corbin, and C.M. D’Antonio. 2009. California native and exotic perennial grasses differ in their response to soil nitrogen, exotic annual grass density, and order of emergence. Plant Ecol. 2001:445–456. Barkworth, M.E., K.M. Capels, S. Long, and M.B. Piep (ed.) 2003. Flora of North America, north of Mexico, vol. 25, Magnoliophyta: Commelinidae (in part): Poacea, part 2. Oxford Univ. Press, New York, NY. Chivers, I.H., and D.E. Aldous. 2005. Reproductive and vegetative responses of different accessions of Microlaena stipoides (Labill.) R. Br. to nitrogen applications and supplementary irrigation in southern Australia. Rangeland J. 27:151–157. Correl, D.S., and M.C. Johnston. 1996. Manual of the vascular plants of Texas. The University of Texas at Dallas, Richardson, TX. Eghball, B., and G.E. Varvel. 1997. Fractal analysis of temporal yield variability of crops sequences: Implications for site-specific management. Agron. J. 89:851–855. Fisher, M.J. 1999. Crop growth and development: Flowering physiology. p. 81–92. In D.S. Loch, and J.E. Ferguson (ed.) Forage seed production, vol. 2: Tropical and subtropical species. CABI Publishing, Wallingford, UK. Gislum, R., and S.M. Griffith. 2004. Tiller production and development in perennial ryegrass in relation to nitrogen use. J. Plant Nutr. 27:2135–2148. Gould, F.W., and V. Kapadia. 1975. The grasses of Texas. Texas A&M Univ. Press, College Station, TX. Harper-Lore, B.L., and M. Wilson (ed.) 1999. Roadside use of native plants. Federal Highway Administration, Washington, D.C. Hatch, S.L., J.L. Schuster, and D.L. Drawe. 1999. Grasses of the Texas gulf prairies and marshes. Texas A&M Univ. Press, College Station, TX. Humphreys, L.R., and F. Riveros. 1986. Tropical pasture seed production. 3rd ed. FAO plant production and protection paper 8. FAO, Rome. Jank, L., and J.B. Hacker. 2004. Setaria. p. 785–807. In L.E. Moser et al. (ed.) Warm-season (C4) grasses. ASA, CSSA, and SSSA, Madison, WI. Jefferson, P.G., B.E. Coulman, and G.A. Kielly. 2001. Production and quality of irrigated timothy hay in Saskatchewan for export hay markets. Agron. J. 93:910–917. Lamb, J.A., R.H. Dowdy, J.L. Anderson, and G.W. Rehm. 1997. Spatial and temporal stability of corn grain yields. J. Prod. Agric. 10:410–414. Littell, R.C., W.W. Stroup, and R.J. Freund. 2002. SAS for linear models, 4th ed. SAS Institute, Cary, NC. Loch, D.S. 1980. Selection of environment and cropping system for tropical grass seed production. Trop. Grassl. 14:159–168. Loch, D.S., B.G. Cook, and G.L. Harvey. 1999a. Location of seed crops: Grasses. p. 113–128. In D.S. Loch, and J.E. Ferguson (ed.) Forage seed production, vol. 2: Tropical and subtropical species. CABI Publishing, Wallingford, UK. Loch, D.S., S.J. Cook, and G.L. Harvey. 1999b. Crop establishment. p. 153–154. In D.S. Loch and J.E. Ferguson (ed.) Forage


seed production, vol. 2: Tropical and subtropical species. CABI Publishing, Wallingford, UK. Loch, D.S., L. Ramírez Avilés, and G.L. Harvey. 1999c. Crop management: Grasses. p. 163–164. In D.S. Loch and J.E. Ferguson (ed.) Forage seed production, vol. 2: Tropical and subtropical species. CABI Publishing, Wallingford, UK. Moore, K.J., K.J. Boole, and M.A. Sanderson. 2004. Physiology and developmental morphology. p. 179–216. In L.E. Moser et al. (ed.) Warm-season (C4) grasses. ASA, CSSA, and SSSA. Madison, WI. National Climatic Data Center. 2008. COOP select state. http:// www7.ncdc.noaa.gov./IPS/coop/coop.html (verified 8 Sept. 2010). U.S. Dep. of Commerce, Washington, D.C. O’Rourke, N., L. Hatcher, and E.J. Stepanski. 2005. A step-bystep approach to using SAS for univariate and multivariate statistics, 2nd ed. SAS Institute, Cary, NC. SAS Institute. 2004. SAS OnlineDoc 9.1.3. SAS Institute Inc., Cary, NC. Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics: A biometrical approach, 2nd ed. McGraw-Hill, New York, NY. Taub, D.R. 2000. Climate and the U.S. distribution of C4 grass subfamilies and decarboxylation variants of C4 photosynthesis. Am. J. Bot. 87:1211–1215. Texas Department of Agriculture. 2007. Agriculture diversification through nature tourism, opportunities for the rural business and land owner. TDA-RED002C (05/07). http://www.

tda.state.tx.us/vgn/tda/f iles/1848/13247_NatureTourismBook2007.pdf (verified 9 Sept. 2010). Texas Department of Agriculture, Austin, TX. USDA-NRCS. 2007. Catarina blend, bristlegrass, Setaria leucopila (Scribn. & Merr.) K. Schum. and Setaria vulpiseta (Lam.) Roemer & J.A. Schultes. E. “Kika” de la Garza Plant Materials Center – South Texas Natives, Kingsville, TX. USDA-NRCS. 2008a. PLANTS profi le, Setaria vulpiseta (Lam.) Roem. & Schult., plains bristlegrass. Available at http:// plants.usda.gov/java/nameSearch (verified 6 Oct. 2010). USDA–NRCS, Washington, DC. USDA-NRCS. 2008b. Characteristics, conservation plant characteristics, Setaria vulpiseta (Lam.) Roem. & Schult., plains bristlegrass, SEVU2. Available at http://plants.usda.gov/java/ charProfi le?symbol=SEVU2 (verified 6 Oct. 2010). USDA– NRCS, Washington, DC. USDA-NRCS. 2008c. Web soil survey. Available at http://websoilsurvey.nrcs.usda.gov/app/ (verified 9 Sept. 2010). USDA– Natural Resources Conservation Service, Washington, D.C. U.S. Naval Observatory. 2008. Sun or moon rise/set table for one year, Form A – U.S. Cites or Towns. Available at http://

www.usno.navy.mil/USNO/astronomical-applications/ data-services/rs-one-year-us (verified 6 Oct. 2010). Naval Meteorology and Oceanography Command, Gulfport, MS. Young, W.C., III, D.O. Chilcote, and H.W. Younberg. 1999. Spring-applied nitrogen and productivity of cool-season grass seed crops. Agron. J. 91:339–343.

WWW.CROPS.ORG

369