Viability of Cyperus esculentus Seeds and Tubers After Ensiling, Digestion by Cattle, and Manure Storage

Abstract

Cyperus esculentus is an invasive sedge causing high losses in many crops. Prevention is key in minimizing further spread and damage. Propagules (tubers or seeds) may spread via cattle manure. This study examined the effect of ensiling, digestion, and storage in manure on the viability of C. esculentus propagules. Propagules were subjected to five durations (0–16 weeks) in silage maize, seven durations (0–48 h) of ruminal digestion, and five durations of storage (0–16 weeks) in manure (slurry or farmyard), or combinations of previous processes. Afterwards, the viabilities were determined by a germination and tetrazolium test. After 6 weeks in a maize silo, the viability of the propagules was reduced by at least 96%. Incubation for 36 h in the rumen, followed by post-ruminal digestion in vitro, reduced seed viability by 30%. However, for the tubers, no effect was observed. The viability of seeds and tubers was reduced by 90% after 11.5 and 13.7 weeks of incubation in slurry, respectively. Compared with seeds, tubers were less tolerant to 12–24 h of animal digestion, followed by 8 weeks of storage in slurry. Keeping a maize silo closed for at least 6 weeks and maintaining slurry storage for at least 16 weeks are excellent measures to eliminate C. esculentus. For farmers, these preventive measures are relatively easy and cheap to implement compared to the requirements of curative control methods.

1. Introduction

Cyperus esculentus L. (yellow nutsedge) is a very competitive weed, ranked as the 16th most troublesome weed in the world [1]. It appears on the European and Mediterranean Plant Protection Organization (EPPO) list of invasive alien species [2]. The weed is hard to control or eradicate because of its high reproduction rate. Under Western European conditions, Bohren and Wirth [3] and De Cauwer et al. [4] found a reproduction factor of 1:746 and 1:638, respectively. Due to a lack of significantly effective single control measures, the weed is very hard to control. In Belgium, C. esculentus is found in almost all arable crops and has infected over 50,000 hectares of cropland (S. De Ryck, pers. comm.). The control is enforced via integrated pest management (IPM) specifications and obligatory weed control [5]. An integrated weed management (IWM) strategy is necessary and should include the prevention of spread [6]. Indeed, prevention is a cornerstone in minimizing the spread of C. esculentus. Indeed, it can spread via ingrowth from neighboring fields or field margins [3,7] but also via contaminated agricultural machinery, plant material, irrigation water, and manure. Hereafter, we focus on the risk of spreading C. esculentus tubers and seeds via contaminated manure. Provided that propagules found in manure are still viable, manure transport can easily lead to propagule migration over hundreds of kilometers.

Until recently, tubers have mainly been assumed to have the greatest impact on the spread of C. esculentus because, unlike plants developed from mother tubers, seedlings might not be sufficiently vigorous to establish in field conditions [8,9]. However, De Ryck [10] demonstrated successful establishment from seeds under Belgian outdoor conditions. Furthermore, seedlings were surprisingly effective at vegetative reproduction to levels comparable with their counterparts grown from mother tubers. In addition, C. esculentus has the potential to produce a high number of viable seeds, particularly when flowers are cross-pollinated. Studies performed in Maine (USA) [11] and in Belgium [10] elucidated that C. esculentus can easily produce 288 to 1500 viable seeds per inflorescence. As a 1 m² patch may easily contain up to 12 mature inflorescences, seed productions of 3456 to 18,000 viable seeds per m² are not exceptional.

After unsuccessful C. esculentus control, seeds and tubers of surviving plants may end up in the harvested produce and feed. This is particularly risky for fields with poorly sensitive C. esculentus populations that escaped control in fields with silage crops (forage maize, other grasses) and tuber and root crops that are occasionally used as a feed (e.g., fodder beet, potato, sugar beet). In many NW European countries, most fields infested with C. esculentus are grown with silage maize, as maize provides better control options for this weed [10]. However, farmers fear that silage is potentially a source of C. esculentus spread, as unsatisfactory control may lead to production of high numbers of viable seeds that will be ensiled with the forage. In heavily infested fields, C. esculentus shoot density can easily exceed 100 shoots m⁻² [12]. Given the late harvest time (end of September), the strongest shoots have ample time to bolt and produce viable seeds within two weeks after flowering (J. Feys, pers. comm.). The weed also thrives in other silage crops, such as ryegrass (Lolium spp.) and alfalfa (Medicago sativa L.). However, in these crops, seed formation is less probable due to regular mowings. In the case of maize fields, aside from this direct contamination with seeds, the silage pile may also indirectly be contaminated with seeds and tubers via contaminated mud clinging to tires of tractors and silage trailers while unloading or compressing the pile, especially when harvesting conditions are wet and soil bearing capacity is poor. Tuber- and root crop-based feeds may become contaminated through soil clinging to roots (indirect contamination) or by roots with internal formation of tubers after being penetrated by rhizomes (direct contamination).

During ensiling, a solid-state lactic acid fermentation process used to preserve biomass to provide farmers with a reserve of livestock feed [13], weed propagules are subjected to unfavorable conditions such as a low pH (2–4), (temporarily) high temperature, high pressure, and the absence of oxygen [14]. Ensiling renders most of these weed seeds non-viable, but reduction in weed seed viability increases with incubation time [15,16] and varies with ensiling environment and species [17,18,19,20]. According to Hahn et al. [21] ensiling unfolds its full seed reduction potential primarily with non-hard-seeded plant species. However, it has a gap in effectiveness when it comes to reducing the viability of hard-seeded plant species, especially if they have reached full maturity and/or display additional dormancy mechanisms. Elema and Scheepens [22] found large reductions in germination ability of C. esculentus tubers with increasing incubation time and burial depth in maize silage piles. For tubers buried 70 cm deep, germination decreased by 82%, 100%, and 100% after 2, 4, and 12 weeks of incubation, respectively. For tubers buried more superficially (5 cm deep), the reductions were smaller, namely 15%, 76%, and 99%, respectively.

After ensiling, the seeds and tubers are ingested by animals. Similar to ensiling, ingestion by cattle and passage through the gut have been shown to reduce seed viability; however, the extent varied with species [15,17,20,23,24,25]. Unfortunately, there are no reports regarding the effect of ensiling on C. esculentus propagules.

Most seeds enter manure via the digestive tract of animals. Different factors, such as manure type, seed depth, storage duration, and temperature, may all affect the vitality of weed seeds [15,22]. Elema and Scheepens [22] and Edwards and Younger [26] investigated the germinability of weed seeds in cattle slurry and farmyard manure, respectively. The authors observed large reductions in germination ability with increasing storage time, burial depth, and storage temperature. To our knowledge, there are no reports concerning the fate of C. esculentus propagules that are stored in manure. Therefore, the aim of our research is to investigate whether C. esculentus propagules can spread through manure. Hereto, we examined the effect of ensiling, digestion by cattle, and storage in slurry and farmyard manure, alone or in combination, on the seed and tuber viability of C. esculentus populations.

2. Materials and Methods

2.1. Plant Material

All performed experiments started with intact, fresh, dark-stored, unimbibed seeds and tubers. Seeds used in this study were collected in July 2023 from outdoor plants from a maize field in Meulebeke (3°19′ E, 50°58′ N), Belgium. After seed harvest, the seeds were air-cleaned and afterwards stored at 4 °C in the dark until the start of the experiments. The cleaned seeds exhibited a thousand kernel weight of 328.1 mg (as determined on four replicates of 1000 seeds) and an initial viability of 96.1 ± 1.23% (as determined on four replicates of 50 seeds).

The tubers used in this study originated from a clonal population sampled in a maize field in Ham (5°10′ E, 51°06′ N), Belgium, and were produced the year preceding the start of the experiments. Between harvest and the start of the experiments, the tubers were stored in a fridge (5 °C). The average individual tuber weight was 360.0 mg. Further, extremely small or large tubers were not used in the experiments. These are tubers weighing less than 80% or more than 120% of the average weight of 360.0 mg.

For the ensiling, digestion, and manure storage experiments, seed samples were enclosed in fine-mesh (50 μm pore size) nylon bags (2 by 5 cm in size), each containing 55 C. esculentus seeds. The pore size was small enough to contain all the seeds while allowing for the passage of water, gases, and microorganisms through the mesh [27]. Tuber samples were enclosed in fine mesh (0.9 mm × 0.8 mm) nylon bags each containing 25 tubers. All experiments started with intact fresh seeds and tubers.

2.2. Ensiling Experiment

In this study, chopped whole plant maize (nominal particle length of approximately 19 mm) was ensiled, without bacterial inoculants, in simple 3.3 L mini-silos, each consisting of a polyvinyl chloride (PVC) pipe (diameter of 11 cm and height of 35 cm) and two rubber end caps. The top end cap had a special plastic valve which prevented air from entering the container, while fermentation gasses could freely pass out, thus ensuring anaerobic conditions. As high DM content silages undergo a restricted fermentation, meaning little acid is produced and the final pH is relatively high [28], two different silage maize materials, differing in dry matter content, were tested. Half of the silos were filled with chopped whole plant biomass of the maize variety SY Glorius, with a DM content of 40.5% (hereafter named high DM maize), and the other half with chopped biomass of P8888, with a DM content of 34.3% (hereafter named “low DM maize”). The dry matter (DM) content of each silage was determined by drying for 24 h at 75 °C in a fan-forced oven. The bags of propagules were layered in the chopped forage to ensure that each bag was in contact with silage. During the filling of the silo, the experimental unit with seeds and tubers was laid down on top of a bottom layer consisting of 800 g of chopped maize, whereafter the silo was filled up with chopped maize to a total mass of 1.25 and 1.5 kg per silo for the wet and dry silage maize, respectively, and compressed with a hydraulic press to eliminate air from the forage mass. Four ensiling durations were maintained per type of silage material, i.e., 6, 9, 12, and 16 weeks. The mini-silos were placed indoors under room temperatures fluctuating between 15 and 23 °C during ensiling duration. At the end of each duration, the silos were opened to retrieve the nylon bags containing seeds and tubers. The experimental design was a completely randomized design with all combinations of types of silage maize and ensiling duration in four replicates.

The pH of unpacked ensiled maize was measured by dissolving 50 g of maize in 50 mL of demineralized water. The average pH (mean ± SE) at 6 and 16 weeks after propagule burial was 3.89 ± 0.02 and 3.87 ± 0.02 for the high DM maize, and 3.84 ± 0.04 and 3.85 ± 0.01 for the low DM maize, respectively. The pH value of both silages was within the optimal range (3.7–4.0) [29], which indicated that the active phase of ensiling was properly achieved. Organic acids were determined by HPLC, according to the methods of Ohmomo et al. [30]. After 6 and 16 weeks of ensiling, the lactic acid concentration (mean ± SE) was 56.54 ± 4.05 and 49.77 ± 3.05 g kg⁻¹ DM for the high DM maize silage, and 57.04 ± 3.66 and 61.35 ± 2.29 g kg⁻¹ DM for the low DM maize silage, respectively. After 6 and 16 weeks of ensiling, the acetic acid concentration (mean ± SE) was 12.31 ± 0.28 and 10.4 ± 0.3 g kg⁻¹ DM for the high DM silage maize and 11.34 ± 0.37 and 11.06 ± 0.33 g kg⁻¹ DM for the low DM silage maize, respectively.

2.3. Digestion Experiment

For the ruminal digestion of seeds, 48 nylon bags, each filled with 55 C. esculentus seeds, were placed into the rumen of two fistulated Holstein-Friesian cows (second year of lactation). The cows were fed a diet composed of 55% whole-plant maize silage, 38% grass silage, and 7% pressed beet pulp (on DM), supplemented by concentrates. Diets were fed for 10 days prior to commencement of the digestion experiment to ensure that the rumen had adjusted to a standard diet. To keep the seed bags submerged in the rumen fluid, the bags were enclosed in heavy, perforated plastic tubes. To avoid any potential risk of contaminating cow dung, the ruminal digestion of tubers (25 per experimental unit) was simulated in vitro using a Daisy II-incubator (ADII; Ankom Technology Corporation Fairport, NY, USA) with rumen fluids retrieved from the previously mentioned fistulated cows. All bags with seeds and tubers were removed from the rumen after 6, 9, 12, 24, 36, and 48 h of incubation (in sacco for seeds or in vitro for tubers) and washed immediately with tap water. After ruminal digestion, stomach (abomasum) and small intestine digestion were simulated in vitro based on the method utilized by Tilley and Terry [31]. The digestion in the abomasum was simulated by incubating the bags in a pepsin-HCl solution (solution of 1 L, pH 1.9, 0.1 M HCl with 2 g L⁻¹ pepsin) for 1 h at 39 °C. After this digestion step, the nylon bags were recovered, washed with tap water, and placed in a pancreatin-KH₂PO₄ solution (solution of 1 L, 0.5 M KH₂PO₄·H₂O with 3 g L⁻¹ pancreatin) at pH 7.75. This solution was shaken for 3 h at 39 °C to simulate small intestine digestion. Large intestine digestion was deemed to have no detrimental effect on propagule viability [14] and was therefore not simulated. Finally, the nylon bags and tubers were recovered and washed with tap water before the start of the germination tests. During the digestion experiment, the welfare of the rumen cannulated experimental cows was strictly respected.

2.4. Slurry and Farmyard Manure Experiment

A total of 32 nylon bags (two types of manure × four durations × four replicates), each containing 55 seeds and 25 tubers, were individually placed in a high-density polyethylene net bag (45 × 65 cm in size, 12 mm pore size). A total of 16 net bags (four durations × four replicates) were placed at a depth of 40 cm in a farmyard manure heap (ground surface of 1.5 m by 1.5 m and height of 1.2 m), and another 16 net bags (four durations × four replicates) were submerged at a 100 cm depth in cattle slurry for 4, 8, 12, and 16 weeks. To keep the net bag submerged in slurry, two bricks were added. During incubation, the manure temperatures were recorded every 10 min using EL-USB-1 dataloggers (LASCAR, Whiteparish, United Kingdom). Figure 1 provides the 16-week time trajectories of manure temperature measured at a depth of 100 and 40 cm in slurry or farmyard manure, respectively. For slurry, the temperature was only monitored during the first 12 weeks.

2.5. Combined Exposures Experiment

All previously mentioned experiments started with intact propagules. In practice, however, when propagules enter the slurry cellar or the farmyard manure heap, most of them have already been subjected to ensiling and digestion. These ensiled or digested seeds are expected to be more susceptible to the subsequent processes like storage in farmyard manure or slurry. In this experiment, we evaluated the joint effect of digestion and slurry incubation on seed and tuber viability. Nylon bags (16 tuber bags and 16 seed bags) containing propagules were digested in sacco (seeds) or in vitro (tubers) for 12 or 24 h, according to the protocols described in Section 2.3. Half of them (8 tuber bags and 8 seed bags) were subsequently incubated in slurry for 8 weeks. The other half (8 tuber bags and 8 seed bags) did not undergo slurry incubation and were stored in darkness at 5 °C until the evaluation of viability.

2.6. Evaluation of Germination and Viability

At the end of each experiment, the tubers were laid on a Copenhagen germination table. In each germination test and for each type of propagule (seeds, tubers), a set of untreated (control) seeds and tubers (dark-stored at 4 °C) were included as a reference. On the Copenhagen table, a regime of an alternating day/night temperature (26/18 °C) under a 16 h light/8 h dark cycle was selected. This temperature regime was chosen to enhance germination [32,33]. Each replicate consisted of two filter papers (Rotilabo Type 112A, Carl Roth GmbH + Co. KG, Karlsruhe, Germany), each bearing 55 seeds or 25 tubers which were moistened by paper strips (Schleicher & Schuell, Dassel, Germany). Seed and tuber germination was monitored for 56 and 10 days, respectively.

Seeds that did not germinate were subjected to a crush test, followed by a tetrazolium viability test [34,35]. In the crush test, all empty seeds or seeds that were crushed by gentle pressure with a pair of tweezers were classified as unviable. The seeds that remained firm were subjected to the tetrazolium test. The seeds were cut longitudinally, and one-half of each seed section was placed on a filter paper moistened with 4 mL of 1% tetrazolium (2,3,5-triphenyltetrazoliumchloride, UCB, Leuven, Belgium). After 24 h of dark incubation at 20 °C, the seeds were evaluated under a binocular microscope. Seeds were considered viable when their embryos were uniformly stained red. Tubers were deemed to be viable, but dormant, if at least 50% of the cutting surface was stained red. The number of viable propagules was calculated as the sum of the number of germinated propagules in the germination test and the number of viable propagules indicated by the tetrazolium test.

2.7. Statistical Analysis

In our experiment, the differences between germinability and viability of C. esculentus seeds were small (<10%), indicating the absence of dormant seeds. For tubers, the difference was also quite small. Statistical analyses were therefore solely based on viability percentages data.

All data were analyzed in R version 4.1.2 [36]. Data were analyzed separately per type of propagule and experiment and were all analyzed using parametric tests run at the 5% significance level, except for the data from the ensiling experiment. Within each category and fermentation process, combinations were treated as a randomized complete block design with four replicates using ANOVA. Treatments comprised all combinations of seven digestion times (including control) and two cows (digestion experiment), with five storage times (including control) in slurry and farmyard manure (manure storage experiments). Because the homogeneity of variances was not met in the data of the ensiling experiment, and the viabilities obtained were all close to zero, a non-parametric test (Kruskal–Wallis test) was performed.

In the combined exposure experiment, a three-way ANOVA for seeds (seven digestion times × two cows × two process trajectories) and two-way ANOVA for tubers (seven digestion times × two process trajectories) were used to check for interactions. To determine the significant differences between group means, either the Tukey HSD test (for normally distributed data) or the Bonferroni test (for non-normally distributed data) was used. For digested seeds and seeds and tubers stored in slurry, the incubation times causing a 50% and 90% reduction in viability were calculated using the three-parameter Weibull (digestion experiment) and three-parameter logistic regression model (manure experiments), according to the methods of Knezevic et al. [37]. Regression analyses were performed using the drc package in R version 4.1.3 [38].

3. Results

3.1. Ensiling Experiment

A dramatic reduction in seed and tuber viability was observed after ensiling. The viability of C. esculentus propagules was significantly affected by 6 to 16 weeks of incubation in manure but not by the DM content of the silage maize. None of the tubers survived ensiling, regardless of the DM content of the silage maize or the incubation period. For the high DM maize silage, only 0%, 0.9%, and 0.9% of viable seeds remained after 6, 9, and 12 weeks of ensiling, respectively. For the low DM maize silage, the viabilities were 0, 0.9%, and 0%, respectively.

3.2. Digestion Experiment

Tuber viability was not significantly affected by 0 to 48 h of in vitro ruminal incubation, followed by post-ruminal digestion (Figure 2).

Seed viability was significantly affected by ruminal incubation time (p < 0.001) but not by dairy cow (p > 0.05), nor was there a significant interaction between dairy cow and incubation time (p > 0.05). The seeds showed a clear decrease in viability with increasing rumen incubation time, averaged over the two dairy cows (Figure 3). The seeds showed a significant reduction in viability after 36 and 48 h of ruminal incubation, followed by post-ruminal digestion. Seemingly, post-ruminal digestion has no effect on seed viability, as seeds exposed to 3 h of ruminal digestion, followed by post-ruminal digestion, were equally as vital as the untreated seeds (0 h of ruminal digestion). The incubation times required for a 90% reduction in seed viability were 81.38 ± 17.88 h for cow 1 and 99.5 ± 27.01 h for cow 2. These estimates were derived from fitted three-parameter Weibull equations that describe the functional relationship between seed viability relative to the untreated control and the ruminal incubation time. Weibull curves and estimated model parameters are provided in Figure 4 and

Table 1. Estimated model parameters (±SE) and corresponding p-values for the three-parameter Weibull models for seed viability as a function of incubation time in the rumen of two lactating dairy cows.

Cow	Parameter	Estimated Value ± Standard Error	p-Value
1	b	2.50 ± 0.89	0.006733
	d	91.45 ± 2.80	<2.2 × 10−16
	e	58.32 ± 6.44	4.145 × 10−12
2	b	2.02 ± 0.66	0.003304
	d	93.48 ± 2.86	<2.2 × 10−16
	e	65.82 ± 9.71	1.340 × 10−8

, respectively.

3.3. Slurry and Farmyard Manure Experiment

The viability of C. esculentus tubers and seeds was significantly negatively affected by the incubation time in slurry (Figure 5,

Table 2. Estimated model parameters (±SE) and p-values for the three-parameter log-logistic models for tuber and seed viability as a function of incubation time in slurry.

Propagule	Parameter	Estimated Value ± Standard Error	p-Value
Tubers	b	2.05 ± 0.48	0.0006394
	d	82.36 ± 4.94	4.268 × 10−11
	e	4.71 ± 0.60	1.104 × 10−6
Seeds	b	3.31 ± 0.84	0.001315
	d	95.81 ± 6.83	4.950 × 10−10
	e	8.26 ± 0.96	3.404 × 10−7

). The incubation times for a 50, 90, and 99% reduction in tuber viability relative to that of the untreated tubers were 4.71 ± 0.60 weeks, 13.75 ± 2.99 weeks, and 44.26 ± 20.9 weeks, respectively. For seeds ED₅₀-, ED₉₀-, and ED₉₉-, the response times were 8.18 ± 2.36 weeks, 11.53 ± 1.77 weeks, and 31.51 ± 10.1 weeks, respectively. However, contrary to the results for incubation in slurry, incubation in farmyard manure did not significantly affect tuber and seed viabilities (p > 0.05), as shown in Figure 6.

3.4. Combined Exposures Experiment

The combined exposures experiment evaluated the joint effect of digestion and slurry incubation on seed and tuber viability. A shown in

Table 3. The significance of the main effects and two- and three-factor interactions in the three-way ANOVA model for seed viability and tuber viability in the combined exposures experiment.

Propagule	Factor	Df	Sum sq	Mean sq	F-Value	p-Value
Tubers	Type of exposure	1	28,884	28,884	402.980	1.340 × 10−10
	Incubation time	1	38	38	0.533	0.479
	Type of exposure Incubation time	1	13	13	0.178	0.680
Seeds	Type of exposure	1	5278	5278	91.955	2.37 × 10−9
	Incubation time	1	72	72	1.257	0.274
	Cow	1	0	0	0.006	0.939
	Type of exposure Incubation time	1	13	13	0.222	0.642
	Type of exposure: cow	1	17	17	0.293	0.594
	Incubation time: cow	1	88	88	1.541	0.228
	Type of exposure Incubation time: cow	1	54	54	0.944	0.342

, the viability of tubers and seeds was only affected (p < 0.05) by type of exposure (digestion with or without an 8 h slurry incubation), not by incubation time (12 h, 24 h) or dairy cow, nor were there any significant interactions between previous factors. Compared to tubers and seeds that were only exposed to 12–24 h of digestion, propagules exposed to both digestion and 8 weeks of storage in slurry displayed significantly (85 and 26.5 percentage points) lower viabilities, respectively (Figure 7).

4. Discussion

This study confirmed that it is possible for C. esculentus seeds to survive ensiling, digestion by cattle, and manure storage. Cyperus esculentus tubers may survive digestion by cattle and manure storage. Farmers should be cautious when using manure as a fertilizer, because it can introduce C. esculentus into their fields.

None of the C. esculentus tubers survived ensiling, irrespective of forage DM content (34.3–40.5%) and ensiling duration (6–16 weeks). Survival was extremely low for ensiled C. esculentus seeds, irrespective of incubation time and forage DM content; only 5 out of 1760 seeds, the total number of seeds buried in the mini-silos across all incubation times and forage DM contents, survived ensiling. Taking into account an initial seed viability of 85.4%, 6–16 weeks of ensiling accounted for a 99.6% reduction in viability. This high sensitivity to ensiling is probably linked to the absence of C. esculentus seeds with hard impermeable seed coats, which make them more vulnerable to silage acids or enable them to imbibe moisture and germinate in the silage environment, provided that no other forms of dormancy are present. These seedlings may succumb due to the lack of oxygen or the presence of phytotoxic acids. Blackshaw and Rode [17] and Westerman et al. [19] indeed found that resistance to ensiling was positively related to the proportion of seeds with an impermeable or hard seed coat. Despite the high sensitivity of propagules to ensiling, the number of seeds surviving ensiling can still be high in the event that the silage heap contains a high number of fully mature seeds, as can occur when the chopped biomass originates from late-harvested maize fields with a high number of seed-bearing inflorescences. Given an average density of 1 inflorescence per m², a fecundity of 288 viable seeds per single inflorescence [10], and a silage yield of 4.5 kg fresh silage maize per m², about 64,000 seeds per tonne will end up in the maize pile, of which 256 seeds will still be alive after 6 weeks of ensiling. Most likely, the minimum incubation time tested in our study was too long to see an effect of forage DM content on viability. If silage acids are the primary cause of seed damage, we would have expected less seed damage in high DM silages. Further, it should be stressed that our experiments were executed in air-tight, well compressed mini-silos, with propagules buried in the center of the silo. Also, the dry matter content, pH value, and lactic acid concentration of our silages were within the optimal range, i.e., 300–400 g/kg, 3.7–4.0 and 3–6%, respectively [29], which indicates that the active phase of ensiling was properly achieved. This might be a reason for the very low propagule vitalities in our experiment. In addition to storage time, seed survival may also depend on burial depth, which was not tested in our study. It is expected that the ensiling conditions are less harsh (less compaction, less anaerobic, higher pH, lower concentrations of phytotoxic compounds, e.g., lactates, propionates, and acetates) at the surface of a silage pile. Elema and Scheepens [22] indeed found that the viability of C. album seeds ensiled for 4 weeks was reduced by 97% at a depth of 70 cm in a maize silo, but by only 53% at a depth of 5 cm. As shorter incubation times were not tested, the critical incubation time allowing for a complete kill could not be determined. However, farmers should minimally respect the duration of 3 weeks required to stabilize the pH of the maize silage at 4 or below. Elema and Scheepens [22] showed that 18% and 0% of the C. esculentus tubers buried 70 cm deep in a maize silo remained viable after 2 and 4 weeks of incubation, respectively. They also found that the number of surviving C. album seeds buried 70 cm deep in a maize silo doubled when the incubation time was reduced from 4 weeks to 2 weeks. Further, a sufficient compaction and a good coverage of the maize pile is key to create and maintain the anaerobic conditions in the maize pile.

Ruminal digestion followed by in vitro simulation of post-ruminal digestion can significantly reduce the viability of seeds, depending on duration of ruminal incubation. Reductions observed in our study are most likely solely caused by ruminal digestion, as the post-ruminal digestion is deemed to have no negative impact on seed viability, as shown for C. album seeds by Aper et al. [15]. Our results show a limited but significant decrease in seed viability of about 29.2 and 45.5% after 36 h and 48 h of ruminal digestion, respectively. The residence time of feed in the rumen of a dairy cow, commonly between 32.8–42 h, largely depends on the milk productivity of the lactating cow, as well as on the diet, specifically, the particle size and density and the composition/forage quality. The ruminal passage rate of feed is higher for high-yielding cows and at high feed intakes [39]. For high-yielding dairy cows receiving a basal diet containing whole-plant silage and concentrates, as in the present study, ruminal passage rates are commonly between 0.037 and 0.063 h⁻¹, corresponding with ruminal residence times between 15.8 and 27 h [40]. Very fibrous diets have slower ruminal passage rates and longer residence times (retention) than grain/concentrate-rich diets, as feed remains in the reticulorumen until the particles are small enough (ideally 3.5–5 mm) to pass into the omasum. Small, dense particles sink to the bottom (grain, well digested forage) of the reticulorum and will more rapidly pass into the omasum in comparison with the time required for lighter, longer particles (recently eaten forage), which float on top of the rumen fluid [41]. Relative to whole-plant maize silage, C. esculentus seeds exhibit higher particle density (1.19–1.34 g mL⁻¹ versus 0.92 g mL⁻¹) and lower particle size (1 mm versus 12.9 mm). Therefore, the ruminal residence time of seeds in the rumen is estimated to be less than 15.8 h. Hence, only minor effects of ruminal digestion on seed viability are expected. As fermentation rate, pH, and ruminal residence time are determined by the diet of the animal, it would be interesting to quantify the effect of different diets on the reduction in seed viability. Mastication is also expected to cause a reduction in seed viability, but this was not investigated here due to the chosen in sacco methodology.

In contrast with seed viability, tuber viability was not affected by ruminal digestion followed by in vitro simulation, regardless of the duration of ruminal digestion. It remains unclear why tubers resist digestion better than do seeds. This differential plant response may possibly be caused by the methodology used for ruminal digestion. In contrast with seeds, tubers were not digested in sacco using rumen-cannulated cows, but rather in vitro. Indeed, the ruminal digestion of C. album seeds for 24 h significantly reduced germinability in a study by Blackshaw and Rode [17] using rumen-cannulated cows, but digestion did not affect germinability in a study by Van Renterghem et al. [42] using an in vitro simulation of ruminal digestion. On the other hand, feed degradation studies show that the in vitro method is a good proxy for the in sacco method, despite the need for further improvements, as stated by Chaudhry and Mohamed [43]. As for the seeds, possible mastication effects on viability are not considered. The damaging effects of mastication may be more pronounced on tubers than on seeds, given their greater size.

The viability of C. esculents seeds and tubers was significantly reduced after slurry incubation. The longer the propagules stayed in a slurry cellar, the higher the reduction in propagule viability. Incubation times for a 90% reduction in tuber and seed viability were 13.75 and 11.53 weeks, respectively. For a 99% reduction, incubation times were 44.3 and 31.5 weeks, respectively. Furthermore, it cannot be excluded that propagules floating on top of the liquid may show a higher chance of survival than the ones in our study, which were kept submersed in the liquid (100 cm below the surface). In the surface layer, propagules may experience less hostile conditions as a result of the lower slurry temperature and higher oxygen concentration relative to those of the deep layers. In contrast with slurry storage, farmyard manure storage did not impair the viability of the seeds and tubers buried 40 cm deep in the heap. Most likely this is due to the sublethal temperatures (<50 °C) generated and maintained in the manure pile, as shown in Figure 1. Storing farmyard manure in far larger solid heaps than those used in our study may certainly impair propagule viability as a result of prolonged exposure to higher temperatures generated in larger heaps, as shown by Rupende et al. [44].

When seeds and tubers enter the slurry cellar or the farmyard manure heap, most of them have already been subjected to ensiling and digestion. As previously shown, almost no propagules survive 6 weeks of ensiling. However, some of them have only been subjected to digestion, a process with only minor detrimental effects on viability. Fortunately, digested tubers were more susceptible to storage in farmyard manure or slurry than non-digested samples. Ruminal incubation and incubation in slurry indeed showed a synergistic effect on the viability of tubers. A total of 8 weeks of ruminal digestion, followed by 12–24 h incubation in slurry, resulted in a 94% lower viability relative to the viability of untreated tubers or tubers that were only exposed to 12–24 h of ruminal digestion. This reduction in viability exceeds the sum of the reductions obtained after 12–14 h of ruminal digestion (0%) and incubation in slurry (77%) (Figure 7). For tubers, some reduction in dormancy during digestion may have occurred, or could even be expected, and this may have increased tuber susceptibility to slurry incubation. However, for seeds, no such synergistic effect could be observed, as the reduction in viability of seeds after 12–14 h ruminal digestion, followed by an 8-week incubation in slurry (50% relative to untreated seeds), is less than the sum of the reductions obtained after 12–14 h of ruminal digestion (4 to 10%) and incubation in slurry (50%) alone. Thus, overall, the impact from ruminal digestion on the viability of seeds in a slurry cellar is deemed irrelevant for the ruminal residence times expected for C. esculentus seeds (12–24 h), particularly for non-dormant types, such as those used in our study. Previously, Blackshaw and Rode [17], Stanton et al. [25], and Piltz et al. [20] also found, for other species, that seed inactivation by the individual fermentation processes (ensiling, digestion, manure storage) does not necessarily increase when seeds are exposed to them sequentially. Overall, compared to seeds, tubers are more resilient to ruminal incubation, followed by incubation in slurry.

5. Conclusions

Ensiling was more detrimental for C. esculentus tubers than for C. esculentus seeds. For slurry storage, the opposite observation was made. Animal digestion only slightly suppressed the viability of propagules, but caused tubers to be more sensitive to the unfavorable conditions in a slurry cellar. To minimize the further spread of viable seeds and tubers of C. esculentus via manure, we recommend the following: (i) keep silos closed for at least 6 weeks, (ii) avoid the input of intact, non-ensiled propagules in the slurry cellar, and (iii) keep the propagules in the slurry for at least 16 weeks. Scarcely any C. esculentus propagule will survive 6 weeks of ensiling, followed by 16-week storage in slurry. As the survival of propagules is greater in farmyard manure than in slurry, farmyard manure should preferentially be spread on infested fields and slurry on non-infested fields. By reducing the vigor of C. esculentus propagules, ensiling helps to exclude C. esculentus from fields and is therefore considered an ecological weed management option within integrated weed management schemes.

For farmers, these preventive measures are easier and cheaper to implement compared to curative control methods, which are more labor-intensive and need to be repeated during a growing season and (often) over years. However, significant challenges still remain. For example, maintaining 16 weeks of slurry storage is not possible if the capacity of the cellar is too small, or if the cellar is already full when the propagules get into it. Keeping a maize silo closed for 6 weeks is a more common practice, but this might be challenging in the case of feed shortage or inaccurate seasonal planning of livestock feed requirements.

Future research might focus on shorter ensiling durations or other green silage materials and explore the robustness of all the aforementioned fermentation-based processes for reducing the viability of C. esculentus propagules under varying climate conditions and dairy farming systems.