Management Implications
Centaurea species, including Centaurea stoebe ssp. micranthos (spotted knapweed) and Centaurea × moncktonii (meadow knapweed), pose an increasing ecological and economic threat in their introduced North American range. These species are highly adaptable and can displace native plants in a variety of environments. One factor complicating Centaurea control is the persistent seedbank, from which new seedlings may continue to emerge following control of established plants. The effects of seed burial depth on emergence of C. stoebe and C. × moncktonii were tested using a realistic outdoor experiment in which seeds emerged from a pasteurized field soil. Emergence was tracked for 3 yr. Most emergence occurred in the fall immediately after seed burial. Emergence declined with increasing burial depth and was eliminated when seeds were buried at 8-cm depth. Based on these findings, managers could potentially reduce establishment of these Centaurea species by inverting the soil so that seeds are buried below 4-cm depth. More generally, these findings suggest that the long potential life span of Centaurea seeds does not necessarily translate into continued emergence over multiple years in the field if the soil remains undisturbed. This finding underscores the importance of controlling the initial flush of newly emerging seedlings in the year following additions to the seedbank.
Introduction
The genus Centaurea within the family Asteraceae includes numerous species that are introduced and invasive in North America (Lejeune and Seastedt Reference Lejeune and Seastedt2001; Roché and Roché Reference Roché and Roché2021). These species have invaded millions of grassland hectares in North America and have caused major reductions in the productivity of more palatable forages (Lejeune and Seastedt Reference Lejeune and Seastedt2001; Roché and Roché Reference Roché and Roché2021). One such species is spotted knapweed [Centaurea stoebe L. ssp. micranthos (Gugler) Hayek]. This economically harmful rangeland weed capitalizes on nitrogen availability and is not effectively suppressed by neighboring vegetation in North America (Callaway et al. Reference Callaway, Waller, Diaconu, Pal, Collins, Mueller-Schaerer and Maron2011; He et al. Reference He, Montesinos, Thelen and Callaway2012). Another Centaurea species is meadow knapweed (Centaurea × moncktonii C.E. Britton), which is a hybrid of black (Centaurea nigra L.), brown (Centaurea jacea L.), and potentially Vochin (Centaurea nigrescens Willd.) knapweeds (Keil and Ochsmann Reference Keil and Ochsmann2006). These invasive knapweeds can achieve rapid population growth in the northeastern United States (Akin-Fajiye and Gurevitch Reference Akin-Fajiye and Gurevitch2020; Molofsky et al. Reference Molofsky, Thom, Keller and Milbrath2023) as well as other areas of the United States (Emery and Gross Reference Emery and Gross2005; Maines et al. Reference Maines, Knochel and Seastedt2013). Both species produce numerous seeds, approximately 3 to 3.5 mm long in C. stoebe and 2.5 to 3 mm long in C. × moncktonii (Minnesota Wildflowers Reference Wildflowers2025).
Rates of recruitment from the seedbank help drive population dynamics in C. stoebe and C. × moncktonii (Akin-Fajiye and Gurevitch Reference Akin-Fajiye and Gurevitch2020; Emery and Gross Reference Emery and Gross2005; Maines et al. Reference Maines, Knochel and Seastedt2013; Milbrath and Biazzo Reference Milbrath and Biazzo2020; Molofsky et al. Reference Molofsky, Thom, Keller and Milbrath2023). Seed dormancy helps determine environmental requirements for germination, the length of time for which viable, nongerminated seeds are likely to remain in the soil, and the likelihood that seeds will eventually germinate at all (Finch-Savage and Leubner-Metzger Reference Finch-Savage and Leubner-Metzger2006). Seed dormancy profiles and germination requirements vary both among and within Centaurea species in North America (Clements et al. Reference Clements, Harmon and Young2010; DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021; Joley et al. Reference Joley, Maddox, Schoenig and Mackey2003; Nolan and Upadhyaya Reference Nolan and Upadhyaya1988; Pitcairn et al. Reference Pitcairn, Young, Clements and Balciunas2002; Watson and Renney Reference Watson and Renney1974; Young et al. Reference Young, Clements, Pitcairn, Balciunas, Enloe, Turner and Harmon2005). In C. × moncktonii, cold-wet stratification reduced dormancy and thereby increased germination (DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021). However, seeds showed germination polymorphism, that is, variability in different seeds’ requirements for the release of primary dormancy. Some seeds did not require stratification to germinate (DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021). Germination polymorphism has been previously reported in other Centaurea species, including C. stoebe (Nolan and Upadhyaya Reference Nolan and Upadhyaya1988). Warmer temperatures (up to 30:20 C day/night, compared with 15:5 C) and light also stimulated germination of C. × moncktonii (DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021). In C. stoebe, afterripening and cold-wet stratification may both reduce dormancy (Eddleman and Romo Reference Eddleman and Romo1988; Nolan and Upadhyaya Reference Nolan and Upadhyaya1988; Watson and Renney Reference Watson and Renney1974). The degree of dormancy may influence whether C. stoebe germination requires light (Nolan and Upadhyaya Reference Nolan and Upadhyaya1988; Watson and Renney Reference Watson and Renney1974).
An experiment with outdoor germination trays found variable surface germination rates in C. × moncktonii, which germinated in either autumn or spring (Milbrath and Biazzo Reference Milbrath and Biazzo2020). The authors noted that high light availability may have enabled autumn germination under the experimental conditions, whereas existing plants might shade the soil surface during autumn under natural conditions (Milbrath and Biazzo Reference Milbrath and Biazzo2020). The same experiment also tested C. stoebe, finding that germination mostly occurred in autumn. For both species, survival of dormant seeds on the soil surface was very low (Milbrath and Biazzo Reference Milbrath and Biazzo2020). In contrast, buried seeds of C. stoebe could remain viable and dormant in the soil for at least 1 yr and potentially more than 8 yr (Akin-Fajiye and Gurevitch Reference Akin-Fajiye and Gurevitch2020; Davis et al. Reference Davis, Fay, Chicoine and Lacey1993). It seems likely that burial reduces germination rates but preserves the viability of seeds in the seedbank.
The objective of this study was to determine the emergence pattern of seedlings of C. stoebe and C. × moncktonii at four soil burial depths over three growing seasons. It was hypothesized that increasing burial depth would reduce emergence of both species and that emergence would be nearly eliminated by 8-cm depth. Emergence was expected to decline with time across the three growing seasons.
Materials and Methods
A trial was established to test the impact of burial depth on emergence of C. stoebe and C. × moncktonii over three growing seasons. The experiment was conducted in an open field area on the Cornell University campus, Ithaca, NY (42.4508°N, 76.4614°W) from 2018 to 2021.
Seed and Soil Collection
Mature flower heads of C. × moncktonii were mass collected in 2018 from the Mt Pleasant Farm, Cornell University, Ithaca, Tompkins County, NY (42.4619°N, 76.3703°W). Flower heads of C. stoebe were similarly collected from a private property in Owego, Tioga County, NY, USA (42.0858°N, 76.3078°W). Seeds were cleaned and counted into lots of 200 using a seed counter (Seedburo Seed Counter, Seedburo Equipment, Des Plaines, IL). Seeds were stored at room temperature in the laboratory for up to 2 wk before the start of the experiment. Initial seed viability was estimated by cold-wet stratifying three lots of 200 seeds of each species at 4 C for 3 mo. Stratified seeds were germinated in an incubator at a thermoperiod of 25:20 C and a photoperiod of 14:10 h (light/dark). Remaining nongerminated seeds were tested for viability by squeezing the seeds with forceps. Hard seeds were considered viable based on previous assessments with a 1% solution of tetrazolium chloride (98% viable; LRM and JB, unpublished data). In contrast, unfilled or dead seeds readily collapsed. The percentage of viable seeds was used as a correction factor when calculating seedling emergence rates in the field: C. × moncktonii: 83.3%; and C. stoebe: 71.2%.
The soil in which the seeds were buried was collected from the Mt Pleasant Farm to provide the same environment for the two Centaurea species. The soil is a very deep, moderately well drained Mardin series (coarse-loamy, mixed, active, mesic Typic Fragiudepts) channery silt loam with pH of 6.3 to 7 (USDA-NRCS 2025). The soil was screened through 1.2-cm hardware cloth to remove stones and large rhizomes, and it was then steam pasteurized at 82.2 C for at least 30 min to eliminate contaminating seeds (Baker and Roistacher Reference Baker and Roistacher1957). This steam pasteurization might have also reduced the incidence of soilborne pests and pathogens, at least at the beginning of the experiment. However, a pasteurization temperature of 82.2 C would not have completely eliminated all microorganisms or altered the physical or chemical properties of the soil, that is, sterilization (Bunt Reference Bunt1988).
Experimental Design
The factors were species, burial depth, and time. The two Centaurea species (C. × moncktonii and C. stoebe) were planted at four seed burial depths (0, 2, 4, and 8 cm). A preliminary study indicated that seedling emergence was unlikely at depths of 8 cm or greater (LRM and JB, unpublished data). This two-way factorial treatment structure was arranged in a completely randomized design with repeated measures on seedling emergence over time. Each treatment was replicated six times for a total of 48 plots (experimental units) measured over six subseasons.
We acknowledge that this multiyear study was not replicated in time, which could have provided more insight into the consistency of seedling emergence patterns. Environmental conditions such as temperature, precipitation, and soil moisture can vary substantially by year, influencing rates of germination and seedling establishment. Replicating the study in time would therefore help determine whether the observed effects of Centaurea species and burial depth on seedling emergence are consistent or variable depending on environmental context.
The experiment was conducted in plots consisting of PVC tubes (15-cm diameter, 102-cm length) buried vertically with 10 cm of the tube above the soil surface. Holes were cut into the lower half of each tube to enhance drainage. Tubes were spaced on 2-m centers in a 7 by 7 grid, that is, a 12 by 12 m area. The area around the tubes was periodically mowed. The tubes were initially filled with local soil, and the soil in the tubes was allowed to settle for 2 mo. The soil is a deep, moderately well-drained Williamson series (coarse-silty, mixed, active, mesic Typic Fragiudepts) very fine sandy loam that is strongly acidic (USDA-NRCS 2025). After settling, soil was removed from the top 35 cm of each tube and replaced with a 30-cm-deep layer of the pasteurized soil described earlier. Two hundred seeds were sown in each plot in mid-September 2018. Seeds of C. × moncktonii and C. stoebe were sown in separate plots at the appropriate depths by scattering the seeds over the soil surface and covering them with additional pasteurized soil, taking into account settling of the soil based on a pilot study. For the 0-cm-depth treatment, seeds were scattered on the soil surface only. A sock sewn from a sheer fabric was secured over the open end of each tube to prevent slug predation and seed contamination. During the winter months, 1.2-cm hardware cloth was added on top of the sock to prevent seed predation by small mammals. The trial was not irrigated.
Emergence rates were calculated over the following 3 yr by monitoring seedling emergence twice weekly from late September to early November 2018, then weekly from May through early November each year for the remaining duration of the experiment. Seedlings were counted and removed at the root crown with blunt forceps. Cumulative percent emergence (adjusted for initial viability) was calculated for each plot over six time points: fall (September through early November) 2018, spring/summer (May through August) 2019, fall 2019, spring/summer 2020, fall 2020, and spring/summer 2021.
Data Analysis
Two plots were excluded from analyses, one from the destruction of the tube by mowing and the second (0-cm treatment) due to extremely poor emergence. Cumulative emergence in the outlier plot was 12% compared with 48% to 76% in the other five replicates, suggesting a non-experimental issue such as loss of the seeds from the soil surface. We have verified that the plot with low emergence is a significant outlier using Grubbs’s test on the logit-transformed cumulative emergence (Grubbs Reference Grubbs1950).
Results were analyzed using a linear mixed model with a logit-transformed response variable (seedling emergence). The model included repeated measures over time and assumed a compound symmetry covariance structure (PROC MIXED, SAS v. 9.4, SAS Institute, Cary, NC). The fixed effects of species, depth, time, and all interactions were included in a full model. A best-fit model was developed through backward elimination (Montgomery et al. Reference Montgomery, Peck and Vining2021; Quinn and Keough Reference Quinn and Keough2002). At each step, the highest-order, least significant interaction term was removed, and then the model was rerun. The process ended when all remaining interaction terms were significant. This approach resulted in a best-fit model containing the fixed effects of species, depth, time, and the species by depth interaction. Treatment means were compared using Fisher’s protected LSD test and the Bonferroni correction (SAS v. 9.4).
Results and Discussion
Monthly average temperatures for 2018 to 2021 are shown in Table 1 (Northeast Regional Climate Center 2025). Total annual precipitation was 106 cm in 2018, 105 cm in 2019, 93 cm in 2020, and 122 cm in 2021 (Northeast Regional Climate Center 2025).
Table 1. Monthly average temperature (C) for Ithaca, NY, USA (Northeast Regional Climate Center 2025).
Seedling emergence varied by burial depth, with percent emergence (averaged over time) showing some minor variation between C. × moncktonii and C. stoebe (species by depth interaction, F(3, 38) = 5.01, P = 0.005; Table 2; Figure 1). Emergence of seedlings was greatest at 0 cm and generally decreased with increasing burial depth, including no successful emergence at 8 cm (Figure 1). For C. × moncktonii, emergence (raw data corrected for seed viability) was approximately 57% at 0 cm, 28% at 2 cm, 3% at 4 cm, and 0% at 8 cm. For C. stoebe, emergence (raw data corrected for seed viability) was 84% at 0 cm, 11% at 2 cm, 4% at 4 cm, and 0% at 8 cm.
Table 2. Type 3 ANOVA for seedling emergence measured for two Centaurea species, four burial depths, and six time points.
a The fixed effects of species, depth, time, and the species by depth interaction were analyzed using a linear mixed model with a logit-transformed response variable. The model included repeated measures over time and assumed a compound symmetry covariance structure (PROC MIXED, SAS v. 9.4, SAS Institute, Cary, NC).
Figure 1. Seedling emergence of two Centaurea species at four burial depths, averaged over time. Raw data are shown (mean ± SE, n = 36). Bars denoted by the same letter are not different (Fisher’s protected LSD test with Bonferroni correction at α = 0.05, analyses performed on a logit scale).
These findings contrast with some studies in which surface-sown seeds had reduced emergence compared with other shallowly buried seeds, typically in larger-seeded species (DiTommaso et al. Reference DiTommaso, Milbrath, Morris, Mohler and Biazzo2017; Froud-Williams et al. Reference Froud-Williams, Chancellor and Drennan1984). In diffuse knapweed (Centaurea diffusa Lam.), shallow burial (0.5 cm) increased emergence relative to surface sowing (Meiman et al. Reference Meiman, Redente and Paschke2009). However, our finding is consistent with a study on lesser starthistle (Centaurea diluta Aiton) that reported a 54% reduction in emergence when seeds were buried at 2 cm compared with 0 cm (Sousa-Ortega et al. Reference Sousa-Ortega, Alcántara, Leon, Barranco-Elena and Saavedra2024). Successful seedling emergence typically declines as seeds are more deeply buried (DiTommaso et al. Reference DiTommaso, Milbrath, Morris, Mohler and Biazzo2017; Froud-Williams et al. Reference Froud-Williams, Chancellor and Drennan1984; Roberts and Chancellor Reference Roberts and Chancellor1979; Roberts and Feast Reference Roberts and Feast1972). Seedlings of species that possess larger-sized seeds can emerge from deeper depths than smaller-seeded species (DiTommaso et al. Reference DiTommaso, Milbrath, Morris, Mohler and Biazzo2017; Froud-Williams et al. Reference Froud-Williams, Chancellor and Drennan1984). Because the seeds of the two Centaurea species we tested are similar in size (average 1.9 mg per seed; JB, unpublished data), we would not expect an interspecific difference in emergence patterns with burial depth.
Similar to our findings, other studies involving Centaurea species have also observed reduced emergence from deep burial. For instance, a growth chamber experiment found that seedling emergence of C. diluta and cornflower (Centaurea cyanus L.) was reduced by 92% and 90%, respectively, when buried at 9 cm compared with 2 cm (Sousa-Ortega et al. Reference Sousa-Ortega, Leon, Lopez-Martinez and Castro-Valdecantos2023). A greenhouse experiment found that seedling emergence of Iberian starthistle (Centaurea iberica Trevis. ex Spreng) was 78% at the soil surface, 22% at 1-cm burial depth, and 0% at 4-cm burial depth (Nosratti et al. Reference Nosratti, Abbasi, Bagheri and Bromandan2017). Combined with our findings, these results demonstrate that it is possible but atypical for Centaurea species to emerge from below approximately 5-cm burial depth.
Because we did not exhume seeds annually, it is unclear how much fatal germination may be occurring, that is, seedlings dying before reaching the soil surface (e.g., DiTommaso et al. Reference DiTommaso, Milbrath, Morris, Mohler and Biazzo2017). Based on previous research as well as an ongoing companion seedbank study, most buried seeds of Centaurea species appear to remain intact and dormant for several years (Davis et al. Reference Davis, Fay, Chicoine and Lacey1993; LRM, AD, JB, SHM, unpublished data). Deeper burial is known to promote dormancy in various weed species (Benvenuti Reference Benvenuti2003; Benvenuti et al. Reference Benvenuti, Macchia and Miele2001). In Centaurea species, including C. × moncktonii, light limitation reduces germination rates (DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021), likely contributing to low emergence of deeply buried seeds. However, these species also show evidence of germination polymorphism (DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021). This phenomenon might help explain why some studies have reported occasional emergence from deep burial (Sousa-Ortega et al. Reference Sousa-Ortega, Leon, Lopez-Martinez and Castro-Valdecantos2023).
The primary flush of seedlings, averaged over Centaurea species and burial depths, occurred during the first few months of the study in fall 2018 (time; F(5, 225) = 5.73, P < 0.001; Table 2; Figure 2). For C. × moncktonii, percentage emergence in fall 2018 (out of all seeds emerged during the experiment for each burial depth) was approximately 95% at 0 cm, 61% at 2 cm, and 67% at 4 cm. For C. stoebe, percentage emergence in fall 2018 (out of all seeds emerged during the experiment for each burial depth) was approximately 99% at 0 cm, 86% at 2 cm, and 97% at 4 cm. Significant but lesser emergence occurred the following season (spring/summer 2019), but little (spring/summer 2020) to no (all other time periods) emergence was otherwise observed (Figure 2).
Figure 2. Cumulative seedling emergence over time, averaged over Centaurea species and burial depth. Raw data are shown (mean ± SE, n = 48). Points denoted by the same letter are not different (Fisher’s protected LSD test with Bonferroni correction at α = 0.05, analyses performed on a logit scale).
The temporal dynamics observed in our study demonstrate that, in the absence of disturbance, emergence can be highly concentrated in the first season after seeds enter the seedbank. For this reason, focusing exclusively on a seed’s maximum longevity can be misleading. Many Centaurea seeds have a relatively protracted maximum longevity. For instance, buried yellow starthistle (Centaurea solstitialis L.) seeds could remain viable for 6 yr (plumeless) or 10 yr (plumed) (Callihan et al. Reference Callihan, Prather and Northam1993). In C. stoebe, more than 50% of buried seeds remained viable and dormant after 5 yr, and 25% remained viable and dormant after 8 yr (Davis et al. Reference Davis, Fay, Chicoine and Lacey1993). However, many seeds have more limited dormancy (DiTommaso et al. Reference DiTommaso, Milbrath, Marschner, Morris and Westbrook2021) or fall victim to pathogens and predators. Milbrath and Biazzo (Reference Milbrath and Biazzo2020) reported that survival of dormant, shallow-buried C. stoebe and C. × moncktonii seeds was very low across multiple sites in New York State. Even seeds that remain viable may still be losing vigor over time, possibly limiting invasion potential. For instance, C. stoebe seeds were reported to germinate more slowly (lower rate of radicle elongation) after burial for 12.5 mo (Chicoine Reference Chicoine1984).
Overall, these findings demonstrate strong impacts of burial depth on emergence in Centaurea species, as well as the steep decline in emergence over time. Management practices that expose Centaurea seeds to soil-surface or near-surface conditions would tend to stimulate emergence, whereas practices leading to deeper burial in soil or litter would reduce emergence. Our results also suggest that controlling the initial flush of seedlings following seed introduction is a crucial step in the process of Centaurea control, although it does not obviate the need for continued follow-up in subsequent years.
Acknowledgments
We thank Erika Mudrak (Cornell University) for statistical advice. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. The USDA is an equal opportunity provider and employer.
Funding statement
Financial support was provided by the U.S. Department of Agriculture (USDA), Agricultural Research Service project no. 59-8062-5-002.
Competing interests
The authors declare no conflicts of interest.
Open access
