lable at ScienceDirect

Crop Protection 29 (2010) 1232e1240

Contents lists avai

Crop Protection

journal homepage: www.elsevier .com/locate/cropro

Effects of winter flooding on weedy rice (Oryza sativa L.)

S. Fogliatto*, F. Vidotto, A. FerreroDipartimento Agronomia, Selvicoltura e Gestione del Territorio, Università degli Studi di Torino, via L. Da Vinci 44, Grugliasco TO, Italy

a r t i c l e i n f o

Article history:Received 23 July 2009Received in revised form7 July 2010Accepted 7 July 2010

Keywords:Red riceGerminationWinter floodingStorage conditionSeed dormancy

* Corresponding author. Tel.: þ39 0116708897; faxE-mail addresses: silvia.fogliatto@unito.it (S. Foglia

it (F. Vidotto), aldo.ferrero@unito.it (A. Ferrero).

0261-2194/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.cropro.2010.07.007

a b s t r a c t

Weedy rice (Oryza sativa L.), characterised by competitiveness, seed longevity, and dormancy is a trou-blesome weed to rice fields. Furthermore, its close botanical affinity to cultivated rice makes its controlparticularly difficult. However, winter flooding of rice fields can be an efficient technique to controlweedy rice infestation by promoting weed seed decay, animal predation, or germination.

The effects of winter flooding on weedy rice plant and seed densities were assessed via two methods:field study of plant densities and laboratory study of germination behaviour.

In the field experiment, weedy rice plant density decreased following application of winter flooding. Infact, winter flooding resulted in more than a 95% reduction in the number of viable weedy rice seeds onthe soil surface as compared to reductions in the range of 26e77% on fields left dry between rice crops.

The laboratory study showed that weedy rice germinability was affected by storage duration, moisturecondition, and thermal regimen. In general, seed germinability increased with storage duration. Bothawnless and awned populations displayed poor germinability under low temperature seed storage, butdisplayed differences when moisture content varied. At �20 �C, we observed, on average, 58% of seeds tobe non-viable when stored for 248 days in water. At þ5 �C the awnless population showed highergermination percentages following shorter storage durations, particularly water stored seeds. At þ25 �C,the highest germination values were recorded in both populations after dry storage, whereas in water,total germination of the awnless population was inversely related to storage duration. Under typical fieldtemperatures, in dry conditions germination behaviour was intermediate between that displayed atþ5 �C and þ25 �C, while storage in water resulted in a faster dormancy breaking in both populations.

The results suggest that winter flooding can be a useful practice to mitigate weedy rice infestations asit promotes germination already in the autumn, before rigorous winter conditions, and favours the decayof non-germinated seeds under low temperature conditions.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Weedy rice (Oryza sativa L.) is a troublesome weed belonging tothe same genus and species as cultivated rice. Weedy rice infesta-tions have been reported to have spread to 40e75% of the total areaof rice cultivation in Europe (Ferrero, 2003), 40% in Brazil (De Souza,1989), 55% in Senegal (Diallo, 1999), 80% in Cuba (Garcia de la Osaand Rivero, 1999), and 60% in Costa Rica (Fletes, 1999). In Italy,where rice accounts for more than 50% of the total European area,weedy rice infestation became significant mainly after the shift intechnique from rice transplanting to direct sowing. Over the last 25years it has becomemore severe particularly after the cultivation ofweak, semi-dwarf indica type rice varieties (Tarditi and Vercesi,

: þ39 0116708798.tto), francesco.vidotto@unito.

All rights reserved.

1993). Infestations are also caused by commercial rice seed,which can contain weed grains.

Weedy rice shows a wide variability of anatomical, biological,and physiological features (Craigmiles, 1978; Kwon et al., 1992;Tang and Morishima, 1997; Vaughan et al., 2001). Some plants areblackhulled with a purple apex and long awns, showing evidence ofwild traits while others are strawhulled and awnless, mimickingcultivated varieties (Federici et al., 2001). The seed pericarp of mostweedy biotypes is pigmented from varying levels of antocyanins,cathekins, and catheolic tannins (Baldi, 1971), causing internationalliterature to identify weedy plants as “red rice.”

Weedy rice infestations are responsible for significant yieldlosses, which are particularly severe in short varieties and lateplantings (Diarra et al., 1985; Kwon et al., 1991; Shivrain et al.,2009). A density of 40 weedy rice plants m�2 has been shown toresult in a crop yield reduction of 46% in Ariette and 58% in Thai-bonnet varieties (Eleftherohorinos et al., 2002). Themanagement ofweedy rice infestations is much more difficult than that of other

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e1240 1233

rice weeds because of the high biological similarity with cultivatedrice and the extended germination over a long period of ricegrowth.

Chemical control with rice selective herbicides is usually inef-fective against weedy forms, except in herbicide tolerant varietiesobtained through genetic engineering or induced mutation. Themost effective control strategies are those that combine preventive,cultural, mechanical, and chemical means, including crop rotations,rice field flooding to more than 5 cm depth of water, and herbicideapplications (Diarra et al., 1985; Chin, 2001; Ferrero et al., 1999).The characteristics of this plant that make it so problematic are itscompetitive advantages over cultivated rice: fast growth, efficientuse of nutrients, resistance to dry conditions, high tillering ability,early grain shattering, and high seed dormancy (Labrada, 2006;Noldin et al., 2006). Gu et al. (2003, 2005) have found that redrice dormancy is associated with the presence of awns and the redcolour of the pericarp and testa, in addition to that imposed by thecovering lemma and palea tissues.

The amount of dormant seeds that are present in the soilconstitutes the soil seed bank; it can be transient or persistentrelative to the dormancy and longevity characteristics of a species(Baldoni and Benvenuti, 2001). Low winter temperatures causebreaking of transient dormancy in summer annuals while highsummer temperatures may induce an increase in dormancy level(Battla and Benech-Arnold, 2007). In fact, temperature is the mainfactor that controls seed dormancy levels, thus regulating the cyclebetween dormancy and germination (Baskin and Baskin, 1998).Previous studies on weedy rice biology have demonstrated that intemperate areas there are essentially two periods during the yearwhen this species can germinate, autumn until the beginning ofwinter and then in early spring (Teekachunhatean, 1985).

The weedy rice seed bank is depleted as a consequence of seedgermination, pathogen attacks, deep seed burial, and animalpredation by arthropods and rodents, and water birds who playa role in fields flooded over winter (Delouche et al., 2007).

In some rice cultivation areas, particularly in the United States,winter flooding is a widespread practice consisting of flooding ricefields from the start of autumn following rice harvest until thespring before tillage operations. Winter flooding is mainly aimed atmanaging rice straw as an alternative strategy to the burning of riceresidue (van Groenigen et al., 2003). Flooding improves rice strawdecomposition primarily because water birds tear the residue intosmaller pieces, which facilitates its soil contact. The consequentdecrease in crop residue and the increase in decomposition makesnitrogen more available the following spring, which reduces thenumber of tillage passes needed to make a seed bed (Bird et al.,2002; van Diepen et al., 2004; van Groenigen et al., 2003).

Winter rice field flooding provides a substitute habitat for manywetland species, in particular, waterfowl that can feed and nestunder these conditions (Czech and Parsons, 2002; Fasola and Ruiz,1996). Previous studies conducted in the United States have shownthat rice field shallow flooding, at a water depth of 15e20 cm, canenhance the density of many species of water birds (Elphick andOring, 1998). This increased waterfowl foraging activity in winterflooded fields may reduce weed biomass at harvest time (vanGroenigen et al., 2003; Bird et al., 2002). It has been establishedthat, grassweeds are the predominant infestation at harvest time inrice fields, and that winter flooding can cause a consistent reduc-tion in grass weed pressure in the subsequent growing season dueto seed foraging waterfowl (van Groenigen et al., , 2003). However,few studies exist on weed behaviour after winter flooding in ricefields, and there is a dearth of knowledge on rice weed germinationpatterns.

This study undertook evaluation of the effects of rice fieldwinter flooding on weedy rice seeds under both field and

laboratory conditions. The field study aimed to determine theinfluence of winter flooding on weedy rice seeds on the soil’ssurface while the laboratory study undertook assessment ofdifferent storage conditions on weedy rice seed germination,including those occurring in winter flooded fields.

2. Materials and methods

2.1. Field study

The study was conducted during 2005e2008 on two sites(Montonero and Sali Vercellese), located in Vercelli province innorthwest Italy. In each site, the study was undertaken on a farmwhich was chosen because either it was highly representative ofa rice cultivation system and weedy rice infestation, or for its wateravailability during winter. At both farms, rice has been cultivated asa mono-crop for at least 50 years.

At Montonero the experimental design included three paddyfields of approximately 2.5 ha each that were managed differen-tially as described below:

(1) winter flooded where the field had never been winter floodedin the years preceding the experiment (OW1);

(2) winter flooded where the field had been winter floodedstarting in the year preceding the experiment (OW2);

(3) maintained dry during the winter period (DRY).

The three differing management techniques were appliedconsistently to their respective fields during the over winterperiods considered in the experimental three-year period.

The experimental design used at Sali Vercellese included OW1and OW2 treatments only due to farm policy to winter flood allfields. Likewise, each field at this site was managed consistentlyover the three experimental years. As treatment management DRYwas excluded from our work, this study can be regarded asa complementary trial to corroborate results obtained in equivalenttreatments (OW1 and OW2) of the Montonero experiment.

At both sites before each rice harvest, the total number of weedyrice plants was counted in three zones (6� 2 m2) randomly chosenwithin each field.

Ten soil core samples were taken from each zone at twodifferent times after harvest (autumn) and at the end of winterflooding (spring). Each core had a diameter of 12 cm and a depth ofbetween 2 and 4 cm. Each soil core was maintained in its owntightly closed plastic bag and stored in a freezer at�20 �C to protectagainst disaggregation. After defrosting, their integrity had beenwell maintained. Weedy rice and rice seeds were initially removedfrom the surface of the cores and put in paper bags for drying. Afterdrying, the hulls of all seeds were removed using forceps todistinguish weedy rice from rice on the basis of pericarp pigmen-tation. As most of the Italianweedy rice biotypes are pigmented, allwhite pericarp seeds were considered rice varieties and not coun-ted in this assessment. The count of weedy rice seeds present on thesoil surface was carried out separately for each core.

2.2. Laboratory study

This study started in October 2006 by hand collecting about 3 kgof weedy rice seeds of both awnless and awned populationsinfesting rice fields on the Montonero farm. The two populationswere considered separately throughout the study. Seeds were driedin open trays for approximately one week at room temperature andthen tested for their germinability after storage under severalconditions for different periods resulting from combining thefollowing factors:

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e12401234

(1) thermal regimen; four regimens were tested: �20 �C (infreezer), þ5 �C (in refrigerator), þ25 �C (in growing chamber)and field temperature;

(2) moisture condition; seeds were stored either dry (in paperbags, containing 90 seeds each) or immersed in water (500 mlbottles filled with tap water, containing 90 seeds each);

(3) storage duration; a total of 18 storage durations were consid-ered: a first period of 10 days, and then by an increasingnumber of two week periods, up to 248 days of storage.

A total of 144 paper bags and 144 bottles were used in the study(two populations� four thermal regimen� 18 storage durations).At the end of each storage period, a bottle and paper bagwere takenfrom each population and thermal regimen to test seed germina-bility. Germination tests were performed in 9 cm-diameter Petridishes with a single layer of filter paper (Whatman No. 1, 10 cmdiameter) and 5 ml of deionised water. All instruments such asforceps, filter papers, pipette tips, and water were previouslysterilised in an autoclave under a pressure of 105 Pa for 10 min. Atseeding time, the seeds from each paper bag or bottle were sani-tized by soaking them in a 0.23% sodium hypochlorite solution for3 min, then rinsed with deionised water, and then randomlydistributed into four Petri dishes (20 seeds each) under a laminarflow hood. All Petri dishes were maintained at a constant temper-ature of 25 �C with an alternation of light/dark (16/8 h). Thenumber of germinated seeds was assessed daily for 14 daysaccording to the International Rules for Seed Testing (ISTA, 2009)for Oryza sativa L. At the end of the germination test, the number ofdead seeds (soft, discoloured, mouldy, without sign of seedlingformation) was also assessed. We calculated Total germination (GT),as the percentage of germinated seeds by the end of germinationtesting (Chiapusio et al., 1997).

2.3. Statistical analyses

2.3.1. Field studyWeedy rice plant and seed counts were expressed as plant and

seed density per square meter. For weedy rice seed density, thereduction percentage after each winter was also calculated for eachzone by comparing the seed density assessed in spring to thatrecorded in autumn (before flooding). This value providesa measure of the variation in weedy rice superficial seed bank thatoccurred during winter. The effect of winter treatment (flooding ordry conditions) onweedy rice plant and seed density over the yearswas tested using the GLM RepeatedMeasure Analysis of Variance ofthe statistical package SPSS (version 16.0). The analysis was

Table 1Weedy rice plant density and seed reduction percentage during winter in dry and flood

Year Winterseason

Dry Ow

Weedy ricedensity(plants m�2)a

Weedy riceseed reduction(%)b

Weden(pla

Montonero1st 2005e2006 18.1 a 26.6 A 32.52nd 2006e2007 4.8 a 77.7 A 15.03rd 2007e2008 19.3 a 31.3 A 11.1

Sali Vercellese1st 2005e2006 e e 20.02nd 2006e2007 e e 8.93rd 2007e2008 e e 7.9

a Lower case letters refer to comparison of weedy rice plant density among the differedifferent (EMMEANS subcommand of GLM procedure in SPSS, see also text; P� 0.05).

b Upper case letters refer to comparison of weedy rice seed reduction percentage amosignificantly different (EMMEANS subcommand of GLM procedure in SPSS, see also text

performed on weedy rice plant density, considering the year as therepeated factor, while winter treatment and zones nested inwintertreatments were the between-subject factors. The same treatmentstructure was adopted for the seed reduction data analysis.However, no nested factors were considered in this case as it waspossible to calculate just the average seed reduction for each zone.The site was not included as a studied factor, thus two separateanalyses were conducted for Montonero and Sali Vercellese.

At Montonero, the interaction between winter treatment andyear was significant, therefore, the main effects of these two factorswere tested by calculating the estimated marginal means and per-formingpairwise comparisons for all level of combinationsusing theEMMEANS subcommand of SPSS, in which the Sidak method wasapplied to adjust the significance level of the comparisons.

2.3.2. Laboratory studyFor each population and storage condition (thermal regimen

and moisture condition) a regression analysis was carried outbetween the storage duration (independent variable) and totalgermination (dependent variable), using the following logisticmodel with three parameters:

GT ¼ a

1þ�

xx0

�b

where GT is total germination (as described in Section 2.2), a, b, andx0 are the curve parameters, and x is the storage duration (in days).In addition, the confidence interval at 5% and at 95% of the totalvariability was also drawn for each curve. The regression analysiswas performed using the software Sigma Plot 2001.

The equations of the fitted curves were then utilized to calculatethe minimum storage duration (in days) required to obtain GTvalues of 20, 50, or 90% (GT20, GT50, and, GT90, respectively).

3. Results

3.1. Field study

3.1.1. Weedy rice plant densityIn Montonero, weedy rice plant density ranged between

4.8 plantsm�2 (DRY, 2nd year) and 32.5 plantsm�2 (OW1, 1st year)(Table 1). If these extreme values are excluded, the density averagedabout 15.0 plantsm�2. This figure falls in the range of the infesta-tion levels that usually occur in Italian rice fields (Vidotto andFerrero, 2009). In Sali Vercellese, plant density was generally

ed treatments at Montonero and Sali Vercellese.

1 Ow2

edy ricesitynts m�2)

Weedy riceseed reduction(%)

Weedy ricedensity(plants m�2)

Weedy rice seedreduction (%)

b 99.9 B 20.0 a 100.0 Bb 98.6 B 12.3 b 99.4 Bb 100.0 B 9.5 b 97.6 B

b 95.8 10.1 a 99.9b 98.9 4.4 a 97.7b 100.0 2.5 a 99.9

nt treatments in the same year. Means sharing the same letter are not significantly

ng the different treatments in the same year. Means sharing the same letter are not; P� 0.05).

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e1240 1235

lower, ranging between 2.5 plantsm�2 (OW2, 3nd year) and20.0 plantsm�2 (OW1, 1st year).

The statistical analysis showed that there was a year effect,a winter treatment effect, and a zone nested in winter treatmenteffect on the final plant density at both sites (Table 2). AtMontoneroa significant interaction was found between year and wintertreatment and between year and zone, while at Sali Vercellese onlythe interaction of year and zonewas observed. At Sali Vercellese theassumption of the repeated measure analysis of variance was notverified because the Mauchly’s test of sphericity was significant,demonstrating that the variances of the differences among therepeated measurements were significantly dissimilar. For thisreason, the GreenhouseeGeisser correction of the degrees offreedom was applied (Park et al., 2009).

At Montonero, the significant interaction between year andtreatment is most likely due to the fact that plant density variationsacross the years showed a different pattern between DRY treatmentand the two flooded treatments. In particular, the DRY treatmentplant density decreased drastically in the second year and rose inthe third year to values similar to those recorded in the first year. Incontrast, in both OW1 and OW2, plant density showed a steadydecline across years. Considering the winter flooded treatmentsonly, the multiple pairwise comparison test performed on theinteraction between year and winter treatment showed that plantdensity of OW1 was significantly higher than that of OW2 in thefirst year, probably because at that time OW1 had not been previ-ously flooded during winter seasons. No significant differencesbetween OW1 and OW2 were observed in the second and thirdyears of the experiment.

At Sali Vercellese, the multiple comparison showed significantlyhigher values of plant density in OW1 compared to OW2 in all threestudy years. Although, similar to the results obtained in the Mon-tonero site, plant density showed a decreasing trend over the yearsin both treatments.

The significant interaction between the year and the zonesfound at both the Montonero and Sali Vercellese sites indicates thatdifferences in weedy rice plant densities may occur among various

Table 2Repeated measures ANOVA of the effect of the year (within-subject factor) andwinter treatment (between-subject factor) on plant density at Montonero and SaliVercellese.

Source of variation SS DF MS F P

Within-subjectMontoneroYear 4943.09 2 2471.55 70.75 0.00Year�winter treatment 3101.62 4 775.40 22.20 0.00Year� zone (winter treatment) 2465.38 12 205.45 5.88 0.00Error 3143.98 90 34.93

Sali Vercellesea

Year 2079.07 1.52 1367.31 44.23 0.00Year�winter treatment 152.84 1.52 100.52 3.25 0.60Year� zone (winter treatment) 1777.19 6.08 292.19 9.45 0.00Error 1410.04 45.61 30.91

Between-subjectMontoneroIntercept 40,723.04 1 40,723.04 1056.53 0.00Winter treatment 1108.33 2 554.16 14.38 0.00Zone (winter) 2856.00 6 476.00 12.35 0.00Error 1734.48 45 38.54

Sali VercelleseIntercept 8608.70 1 8608.70 333.18 0.00Winter treatment 1160.67 1 1160.67 44.92 0.00Zone (winter) 1774.76 4 443.69 17.17 0.00Error 775.13 30 25.84

a The GreenhouseeGeisser correction of the degree of freedom has been applied.

zones within a paddy field, and that these differences may changeover the years.

3.1.2. Weedy rice seed reductionAt both sites, winter flooding produced consistent reductions of

more than 95.0% (Table 1) in weedy rice seed density on the soilsurface when compared to those recorded in autumn. Over-wintering under conventional dry conditions yielded reductionsranging from 26.6% (first year) to 77.7% (second year).

The repeated measure analysis at Montonero showed not onlya year effect, but also a significant interaction between year andwinter treatment (Table 3). At Sali Vercellese, no differences werefound across years.

In all three years at Montonero, the multiple comparison per-formed on the interaction showed that the DRY treatment causeda significantly lower weedy rice seed winter reduction as comparedto the flooded treatments. Moreover, no differences in seedreduction were found between OW1 and OW2.

At Sali Vercellese, no differences were recorded across the yearsor between the two treatments. In particular, a nearly completespring depletion of weedy rice seed density was observed alreadyafter a year of winter flooding.

3.2. Laboratory study

Our results showed the strong influence of storage conditions(thermal regimen and moisture condition) and storage durationupon the total germination percentage of both weedy rice pop-ulation seeds. At�20 �C, dormancy was almost preserved under allconditions, regardless storage length. Water stored awned seeds(Fig. 1) displayed the highest total germination value (36.0%) whilethe awnless population was unenhanced by water, reachinga maximum of only 7.0% after 10 days of storage. Under dryconditions, it took 234 days of storage to achieve a germinationpeak of 15.0%. The awned population, kept in water, yielded highergermination values with shorter storage compared to dry condi-tions. The fitting of the logistic model to these data was very poorand the low GT values did not allow calculation of GT20, GT50, andGT90. In the case of water storage, a considerable amount ofungerminated seeds were actually dead (average of 51% and 66% inawned and awnless populations, respectively; data not shown).Dead seeds averaged only 2.7% in dry conditions.

Table 3Repeated measures ANOVA of the effect of the year (within-subject factor) andwinter treatment (between-subject factor) on seed reduction at Montonero and SaliVercellese.

Source of variation SS DF MS F P

Within-subjectMontoneroYear 1537.31 2 768.65 26.31 0.00Year�winter treatment 3256.49 4 814.12 27.86 0.00Error 350.63 12 29.22

Sali VercelleseYear 14.28 2 7.14 0.54 0.60Year�winter treatment 24.05 2 12.03 0.92 0.44Error 105.08 8 13.13

Between-subjectMontoneroIntercept 178,214.93 1 178,214.93 9369.18 0.00Winter treatment 17,519.99 2 8759.99 460.53 0.00Error 114.13 6 19.02

Sali VercelleseIntercept 175,403.72 1 175,403.72 18,303.86 0.00Winter treatment 3.67 1 3.67 0.38 0.57Error 38.33 4 9.58

Fig. 1. Total germination percentage of awnless and awned weedy rice seeds stored for different durations at a constant �20 �C in water or dry conditions.

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e12401236

After constant þ5 �C storage, the awnless population presenteda higher total germination compared to the awned one, whetherseeds were immersed in water or kept in dry conditions (Fig. 2). Inthe awnless population stored in water, GT increased to about 95%during thefirst 136days of storage and showed a slight reduction forlonger storage durations. Total germination of awned populationpeaked at 73.0% after 248 days of storage in dry conditions. Anincreasing trend ofGTover storage durationwas particularly true forthe awnless population; it was also evident that water storageresulted in a faster seed dormancy release relative to dry storage.Thiswas confirmed byGT20 andGT50 values, whichwere 21 days and56 days in water, and 70 days and 155 days in dry conditions,respectively. The awned population exhibited a marked dormancy,particularly in the case of water storage. In this case, GT did notexceed30%.When the awned seedswere stored indryconditions,GTwas initially low, but started to rise after about 200 days of storage,giving GT20 and GT50 values of 193 days and 238 days, respectively.Even though GT values higher than 90% were recorded in someawnless water stored populations, the fitted model was unable toestimate the value of GT90 within the maximum storage durationconsidered (GT90> 248 days). Averaging among storage durationsandpopulations, thedead seeds at the endof germination testswere3.3% and 4.8% in dry and water conditions, respectively.

At a constant þ25 �C, the highest GT values were recorded inseeds stored under dry conditions for both the awnless and awnedpopulations (97.0% after 248 storage days) (Fig. 3). A particulartrend was exhibited in the case of the water stored awnless

population as GT was generally inversely related to storage dura-tion. In fact, the highest GT value (73.0%) was obtained from seedsstored for 10 days, while the lowest values were reached after 200days of storage. The same population stored under dry conditionsshowed opposite behaviour, reaching GT values of 32.0% after only10 days of storage and 100.0% at 38 days and beyond.

The awned population stored in dry conditions behaved, on thewhole, similarly to that displayed by the awnless one maintained inthe same storage condition; it reached a GT value of 97.0% after 248storage days. The water stored awned population showed more vari-able and lower GT values, resulting in a GT20 of 138 days. Overall, theamount of dead seeds at the end of germination tests was 8.2% and14.1% in dry and water conditions, respectively when storage dura-tions and populations were averaged.

Seeds stored under field thermal regimen produced GT valuesgreater than 90% in both populations and at the two moisture levels(Fig.4). Inbothpopulations, the storageduration required to reachthehighestGT valueswas shorter in seeds stored inwater. Inparticular, inthe awnless population stored in water, GT achieved about 40% afterjust 10daysof storage, and95%afterabout120dayswhile seedsof thesamepopulationmaintained indrycondition requiredabout180daysof storage to reach GT values exceeding 90%. Initially, a low germina-tion percentagewas recorded for the awned populationdwhether inwaterordry storage conditionsdeven if the fastest increase inGTwasagain demonstrated by seeds stored inwater.

In general, a larger variabilityofGT valueswasobtained fromseedsstored inwater rather than observed in dry conditions, which did not

Fig. 2. Total germination percentage of awnless and awned weedy rice seeds stored for different durations at a constant þ5 �C in water or dry conditions.

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e1240 1237

allow us to estimate the value of GT90 within 248 days, even thoughseveral GT values exceeded 90%. Averaging among storage durationsand populations, the amount of dead seeds at the end of germinationtests was 2.7% and 3.1% in dry and water conditions, respectively.

4. Discussion

Study results highlight the significant efficacy of winter floodingin reducing weedy rice infestations in paddy fields. In particular,this practice showed a remarkable reduction of weedy rice seeddensity in comparison to overwintering under dry conditions.

Weedy rice plant density showed a decreasing tendency overthe years in all fields that were kept flooded during the winter; weobserved more variation in behaviour in the field that underwentconventional overwintering in dry condition. The strong reductionof the weedy rice seed bank on the soil surface in all conditionsafter winter floodingmay be the reason for the gradual reduction ofthe weedy rice plant density. However, it should be noted that welimited this study to the effects of winter field conditions on theseeds that dropped on the soil surface during the last crop season.Weedy rice plants could also have resulted from seeds that hadbeen produced in previous years and already incorporated into thesoil. This may explain why reductions in the density of weedy riceplants were weaker than those of seed density.

The study showed that even a single winter of field flooding canbe very effective in reducing seed density, and that multiple yearrepeatedwinterflooding eventually results in low infestation levels.

The results of the laboratory study indicated that both temper-atures and moisture storage conditions influenced weedy ricegermination. Water storage induced an anticipation of the germi-nation, particularly in the awnless population atþ5 �C,þ25 �C, andat the field thermal regimen. Generally, the awned populationshowed a high degree of dormancy, demonstrated by the lowvaluesof total germinationand the longperiodof storage required toobtainany significant germination. This behaviour can be explained by thefindings of Gu et al. (2005) that revealed the existence of a directcorrelation between dormancy and the awn presence.

Storage temperatures had different effects on germination. Afterstorage at �20 �C, the ability of seeds to germinate was almostcompletely hampered, as found in other studies of storage tempera-tures at�20 �C (Cohn and Hughes, 1981), �15 �C (Teekachunhatean,1985), and �10 �C (Agostinetto et al., 2001). In addition, in theseconditions, the germination was also very low because of deteriora-tion phenomena that occurred during storage. In particular, a highamountof dead seeds at the endof germination tests (about 66%)wasobserved in the awnless population stored inwater.

At þ5 �C, the awnless population in water germinated fasterthan the awned one, and a stimulation of the germination processwas evident after about 190 days in dry conditions. These resultslikely meant that the temperature around þ5 �C can favourdormancy breaking as previously observed by Gianinetti and Cohn(2008), particularly in awned populations.

An unexpected germination response was exhibited by theawnless population stored in water at þ25 �C, in which initial

Fig. 3. Total germination percentage of awnless and awned weedy rice seeds stored for different durations at a constant þ25 �C in water or dry conditions.

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e12401238

germination was high (73%) and then decreased over the storageduration. This behaviour can likely be related to the relatively highpercentage of dead seeds recorded at the end of the experiment,and the induction or maintenance of the dormant status ina number of seeds over time. These assumptions are consistentwith the findings of Teekachunhatean (1985), which pointed outthat seeds stored at room and high temperatures in water wereprone to deterioration and likely to become dormant. In particular,the þ25 �C temperature might have induced secondary dormancyover the storage period (Gianinetti and Cohn, 2008). Generally, theeffect of thermal regimen was influenced by moisture condition.Under field regimen, for example, the behaviour of total germina-tion in relation to storage duration in dry condition was interme-diate between that displayed at þ5 �C and þ25 �C while this wasnot evident in the case of storage in water.

The germination behaviour of weedy rice after storage underdiffering conditions could partially explain the results of the fieldstudyandpermit us tomake several hypotheses about seeddepletionafter winter flooding. First, under flooding conditions, a considerableamount of seeds can deteriorate through low and high temperatureexposure, particularly to temperatures below zero, as observed in thecase of storage in water at �20 �C. As reported by Nelms and Twedt(1996), winter flooding can favour the decay of seeds of severalweeds including weedy rice, even if it is less prone to degradationthan other weed and crop seeds because of its hulls.

Second, while some seeds germinate in autumn, just afterflooding begins, others germinate at the end of the flooding period

when dry conditions are re-established. A reduction in the weedyrice seed bank caused by these germination processes might beenhanced by the presence of water. Seeds that germinate inautumn are destroyed by winter low temperatures (suicidalgermination), whereas the seeds that germinate in spring geteliminated by tillage operations performed before crop seeding. Inaddition, storage in water under field conditions was particularlyeffective in promoting dormancy release in the awnless population,which is by far the most abundant in the area of the field study.

Several other phenomena may have combined to cause weedyrice seed reduction during winter, including seed predation. Manystudies have asserted that flooded harvested rice fields can attractwaterfowl, which feed on weed seeds in winter; moreover, weedyrice seeds are larger than those of many other weeds and seem tobe preferred by birds (Holland et al., 2008).

Winter flooding can be regarded as a fundamental method ofweedy rice control both in organic farming and in sustainablesystems. In fact, winter flooding starting in the post harvest autumnperiod can enhance weedy rice seed germination, in a way similarto a stale seed bed intervention. With winter flooding, weedy riceseedlings are mostly controlled by low winter temperatures, thusthey do not need to be sprayed with broad spectrum herbicides,such as glyphosate (Valverde, 2005), as commonly required withthe conventional stale seed bed.

Winter flooding can reduce the amount of weedy rice seedemergence in the spring, and reduce the level of weedy riceinfestation at crop establishment. Yet, it cannot be considered an

Fig. 4. Total germination percentage of awnless and awned weedy rice seeds stored for different durations under field thermal regimen in water or dry conditions.

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e1240 1239

absolute means of weedy rice infestation control; herbicide appli-cation may still be required against further flushes of emergence. Itis reasonable to believe that winter flooding could be profitable inmost of the Italian rice area without causing any drastic alterationof common rice cultivation practices, including varietal choice,tillage procedures, weed control, and water management duringrice cultivation. The most significant limiting factor is water avail-ability during the winter.

Theoretically, winter flooding, as well as stale seed bed, couldlead to selection of weedy rice populations with deeper dormancy,although this phenomenon has yet to be reported (Vidotto andFerrero, 2005). Nonetheless, one must consider that the level ofinfestation in subsequent seasons will depend on the differentweedy rice populations that comprise the weed community ofa field. In the presence of blackhull or awned populations, it is likelythat their seed dormancy will last a long time (Gu et al., 2003). Or,when a strawhull population is dominant, as in the case of thisstudy, dormancy will likely be quickly overcome.

Further studies is needed to assess the relative importance ofthe different causes of seed depletion after winter flooding, and tobetter understand the germination behaviour of weedy rice.

Acknowledgements

The authors express their gratitude to all members of the weedresearch group of the Department of Agronomy, Forest and LandManagementof theUniversityof Torino for theirassistance in thefield

and laboratory works. The authors also want to recognize the anon-ymous reviewers and the Editor for their valuable contribution to theimprovement of this paper. The paper is attributable in equal parts tothe authors.

References

Agostinetto, D., Fleck, N.G., Costa, E.L.N., Vidal, R.A., Merotto Júnior, A., 2001.Liberação da dormência em arroz vermelho: ações do período e da tem-peratura de armazenamento e da integridade física das sementes. ScientiaAgraria 2, 17e23.

Baldi, G., 1971. Presenza del carattere Pericarpo rosso in varietà di riso coltivate(O. sativa). Il riso 20, 299e302.

Baldoni, G., Benvenuti, S., 2001. Dispersione dei propagali nel tempo. In:Catizone, P., Zanin, G. (Eds.), Malerbologia. Pàtron editore, Bologna, pp. 110e111.

Baskin, C.C., Baskin, J.M., 1998. Seeds: Ecology, Biogeography, and Evolution ofDormancy and Germination. Academic Press, San Diego.

Battla, D., Benech-Arnold, R.L., 2007. Predicting changes in dormancy level in weedseed soil banks: implications for weed management. Crop Prot. 26, 189e197.

Bird, J.A., Eagle, A.J., Horwath, W.R., Hair, M.W., Zilbert, E.E., van Kessel, C., 2002.Long term studies find benefits, challenges in alternative rice straw manage-ment. Calif. Agric. 56, 69e75.

Chiapusio, G., Sanchez, A.M., Reigosa, M.J., Gonzalez, L., Pellissier, F., 1997. Dogermination indices adequately reflect allelochemical effects on the germina-tion process? J. Chem. Ecol. 23, 2445e2453.

Chin, D., 2001. Biology and management of barnyardgrass, red sprangletop andweedy rice. Weed Biol. Manag. 1, 37e41.

Cohn, M.A., Hughes, J.A., 1981. Seed dormancy in red rice (Oryza sativa) I. Effect oftemperature on dry-afterripening. Weed Sci. 29, 402e404.

Craigmiles, J.P., 1978. Introduction. In: Eastin, E.F., Red Rice Research and Control,Tex. Agric. Exp. Stn. Bull., pp 5e6.

Czech, H.A., Parsons, K.C., 2002. Agricultural wetlands and waterbirds: a review.Waterbirds 25 (Special Publication 2), 56e65.

S. Fogliatto et al. / Crop Protection 29 (2010) 1232e12401240

Delouche, J.C., Burgos, N.R., Gealy, D.R., Zorilla de San Martín, G., Labrada, R.,Larinde, M., Rosell, C., 2007. Weedy Rices e Origin, Biology, Ecology and Control.In: FAO Plant Production and Protection Paper, vol. 188 85e86.

DeSouza,P.R.,1989.Arrozvermelho:umgrandeproblema.LavouraArrozeira42, 30e31.Diallo, S.,1999.Problèmeposépar le riz rougeenrizicultureauSénégal.Reportof theglobal

workshop on red rice control, 30 Auguste3 September 1999, Varadero, Cuba, 45e49.Diarra, A.R.J., Smith, R.J., Talbert, R.E., 1985. Red rice (Oryza sativa) control in drill

seeded rice (O. sativa). Weed Sci. 33, 703e707.Eleftherohorinos, I.G., Dhima, K.V., Vasilakoglou, I.B., 2002. Interference of red rice

in rice grown in Greece. Weed Sci. 50, 167e172.Elphick, C.S., Oring, L.W., 1998. Winter management of Californian rice fields for

waterbirds. J. Appl. Ecol. 35, 95e108.Fasola, M., Ruiz, X., 1996. The value of rice fields as substitutes for natural wetlands

for waterbirds in the Mediterranean region. Colonial Waterbirds 19 (SpecialPublication 1), 122e128.

Federici, M.T., Vaughan, D., Tomooka, N., Kaga, A., Wang, X.W., Doi, K., Francis, M.,Zorilla, G., Saldain, N., 2001. Analysis of Uruguayan weedy rice genetic diversityusing AFLP molecular markers. Electron. J. Biotechnol. 4, 1e16.

Ferrero,A.,Vidotto,F.,Balsari,P.,Airoldi,G.,1999.Mechanicalandchemicalcontrolof redrice(Oryza sativa L. var. sylvatica) in rice (Oryza sativa) pre-planting. CropProt.18, 245e251.

Ferrero, A., 2003. Weedy rice, biological features and control. Weed Managementfor Developing Countries. In: FAO Plant Production and Protection Paper, vol.120, Addendum 1, pp. 89e107.

Fletes, M.S., 1999. Evaluation de la maleza arroz rojo (Oryza sativa) en las principaleszones arroceras de Nicaragua. Report of the global workshop on red ricecontrol, 30 Auguste3 September, Varadero, Cuba, 41e44.

Garcia de la Osa, J., Rivero, L.E., 1999. El arroz rojo. Estudios y perspectivas de sumanejo en la producción arrocera cubana. Report of the global workshop on redrice control, 30 Auguste3 September, Varadero, Cuba, 25e31.

Gianinetti, A., Cohn, M.A., 2008. Seed dormancy in red rice. XIII: interaction of dry-afterripening and hydration temperature. Seed Sci. Res. 18, 151e159.

Gu, X.Y., Chen, Z.X., Foley, M.E., 2003. Inheritance of seed dormancy in weedy rice.Crop Sci. 43, 835e843.

Gu, X.Y., Kianian, S.F., Foley, M.E., 2005. Seed dormancy imposed by covering tissuesinterrelates to shattering and seed morphological characteristics in weedy rice.Crop Sci. 45, 948e955.

Holland, J.M., Smith, B.M., Southway, S.E., Birkett, T.C., Aebischer, N.J., 2008. Theeffect of crop, cultivation and seed addition for birds on surface weed seeddensities in arable crops during winter. Weed Res. 48, 503e511.

ISTA (International Seed Testing Association), 2009. International Rules for SeedTesting, pp. 2e18 Bassersdorf, Switzerland.

Kwon, S.L., Smith Jr., R.J., Talbert, R.E., 1991. Interference of red rice (Oryza sativa)densities in rice (Oryza sativa). Weed Sci. 39, 169e174.

Kwon, S.L., Smith Jr., R.J., Talbert, R.E., 1992. Comparative growth and developmentof red rice (Oryza sativa) and rice (O. sativa). Weed Sci. 40, 57e62.

Labrada, R., 2006. Weedy rice problems and solutions for its management. In:International Rice Commission Newsletter, vol. 55. FAO, Rome, pp. 114e118.

Nelms, C.O., Twedt, D.J., 1996. Seed deterioration in flooded agricultural fieldsduring winter. Wildl. Soc. Bull. 24, 85e88.

Noldin, J.A., Chandler, J.M., McCauley, G.N., 2006. Seed longevity of red rice ecotypesburied in soil. Planta Daninha 24, 611e620. Viçosa-MG.

Park, E., Cho, M., Ki, C.S., 2009. Correct use of repeated measure analysis of variance.Korean J. Lab. Med. 29, 1e9.

Shivrain, V.K., Burgos, N.R., Gealy, D.R., Smith, K.L., Scott, R.C., Mauromoustakos, A.,Black, H., 2009. Red rice (Oryza sativa) emergence characteristics and influenceon rice yield at different planting dates. Weed Sci. 57, 57e94.

Tang, L., Morishima, H., 1997. Genetic characterization of weedy rices and theinference on their origins. Breed. Sci. 47, 153e160.

Tarditi, N., Vercesi, B., 1993. Il riso crodo: un problema sempre più attuale in ris-icoltura. L’Informatore Agrario 11, 91e95.

Teekachunhatean, T., 1985. Release, induction and significance of dormancy in seedsof red rice (Oryza sativa L.), Mississippi State University, USA. (PhD dissertation)

Valverde, B., 2005. Can feral rice remain undetected? In: Gressel, J. (Ed.), CropFerality and Volunteerisms. CRC Press, Boca Raton, pp. 279e294.

van Diepen, L.T.A., van Groenigen, J.W., van Kessel, C., 2004. Isotopic evidence forchanges in residue decomposition and N-cycling in winter flooded rice fields byforaging waterfowl. Agric. Ecosyst. Environ. 102, 41e47.

van Groenigen, J.W., Burns, E.G., Eadie, G.M., Horwath, W.R., van Kessel, C., 2003.Effects of foraging waterfowl in winter flooded rice fields on weed stress andresidue decomposition. Agric. Ecosyst. Environ. 95, 289e296.

Vaughan, L.K., Ottis, B.V., Prazak-Havey, A.M., Bormas, C.A., Sneller, C., Chandler, J.M.,2001. Is all red rice found in commercial rice really Oryza sativa? Weed Sci. 49,468e476.

Vidotto, F., Ferrero, A., 2005. Modelling population dynamics to overcome feral ricein rice. In: Gressel, J. (Ed.), Crop Ferality and Volunteerisms. CRC Press, BocaRaton, pp. 355e370.

Vidotto, F., Ferrero, A., 2009. Interactions between weedy rice and cultivated rice inItaly. Ital. J. Agron. 4, 127e136.