Responses of Phalaris minor Rezt. and Phalaris brachystachys Link to different levels of soil water availability

Phalaris brachystachys and Phalaris minor are common and troublesome weeds in winter cereals in Mediterranean countries. Different distribution and soil preferences have been found for each species in Andalusia (southern Spain). In irrigated fields P. minor is more frequent while P. brachystachys has extended its range to semiarid provinces with low rainfall. This different adaptation to irrigation conditions is difficult to explain considering aspects of their biology, herbicide tolerance, or cultivation practices. The objective of this study was to assess the influence of different soil water availabilities over growth and reproductive aspects to explain the differences found in ecology and distribution of P. brachystachys and P. minor . The experiment was conducted under greenhouse controlled conditions using five levels of water availability: field capacity, light drought, moderate drought, severe drought and extreme drought. Differences between species and among treatments were found in plant height, biomass, tiller number, and reproductive traits. Field capacity and light drought treatments favoured biomass, tiller number, and panicle number in P. minor . In contrast, P. brachystachys had a positive response only in moderate drought and increased the percentage of mature panicle with increasing drought levels. These results could explain the wider distribution of P. brachystachys in fields without supplemental irrigation in semiarid areas, due to its adaptation to moderate drought conditions. It may also clarify the greater frequency of P. minor in irrigated fields and in areas with higher rainfall.

In Spain these species infest crops throughout autumn and winter, creating a serious problem for cereal production (Pujadas-Salvá, 1986;Saavedra et al., 1989b;Hidalgo et al., 1990). Due to the high cost and difficulty of control measures (González-Díaz et al., 2009), as well as their high rate of population growth (González-Andújar et al., 2005), farmers consider canary grasses some of the most pernicious weeds of cereal crops. In fact, several studies have indicated that competition between canary grass species and winter cereals is severe (Cudney and Hill, 1979;Dellow and Milne, 1986;Mehra and Gill, 1988;Walker et al., 2001), consequently sustainable weed management practices are being exploited as control strategies (Franke et al., 2007;Jamil et al., 2009).
In Andalucia (southern Spain) P. brachystachys is the most frequent and extended of the canary grass species (Saavedra et al., 1989a;González-Andújar and Saavedra, 2003); however, P. minor is also widely distributed and produces more severe infestations than P. brachystachys. In contrast, P. paradoxa produces the most highly-infested areas in terms of field density (Saavedra et al., 1989a), in spite of its reduced distribution. Differences in P. paradoxa's germination capacity, such us greater light dependence (Jiménez-Hidalgo et al., 1997;Taylor et al., 2004) and a more extended period of emergence (Jiménez-Hidalgo, 1993;Mancebo et al., 2007), may explain its reduced and distinct distribution compared to the other species.
To our knowledge, neither differences in biology nor competition between P. minor and P. brachystachys have been found. In fact, the competitive abilities of P. minor and P. brachystachys are similar in Andalusia (Jiménez-Hidalgo et al., 1997) and for their biotypes in Greece (Afentouli and Eleftherohorinos, 1999). However, specific soil preferences have been found for each species in Andalusia. In irrigated fields, P. minor is more frequent in loamy soils, while P. brachystachys appears to prefer heavy clay soils (Saavedra, 1987;Saavedra et al., 1990). In dry land fields, however, there were no differences in frequency (Hidalgo, 1988;Hidalgo et al., 1990) between the species. Nevertheless, P. brachystachys has extended its range to semiarid provinces with low rainfall (Saavedra et al., 1989a).
This different adaptation to soil and irrigation conditions is difficult to explain with aspects of their biology, herbicide tolerance, or cultivation practices. However, they may indicate a different tolerance to deficient or excessive watering regimens. Other authors (Rodiyati et al., 2005) have related differences in distribution of Cyperus species to drought tolerance. In Phalaris species, the more extended distribution of P. brachystachys could be linked with drought tolerance, but to our knowledge, has not yet been established. Hence, the objective of this study was to assess the influence of different soil water levels over growth and reproductive aspects under greenhouse controlled conditions to explain the differences found in distribution of P. brachystachys and P. minor.

Material and methods
The experiment was conducted in a greenhouse in Alameda (Córdoba) to evaluate drought tolerance in 1988, from January to May. Temperatures during development, from February to May, ranged from 20 to 30°C at day and from 10 to 18°C at night. Plants were exposed to 14 hours of light and 10 of darkness.
P. brachystachys and P. minor seeds were previously collected in June in cultivated fields in Córdoba, identified, confirmed (Tutin, 1980), and stored in glass containers in the laboratory during summer and autumn. Ten seeds per species were sown on January 22 in 15 cm diameter pots with 1,700 g of substrate composed in volume of three equal parts of sand, silt, and turf plus 1 g of fertilizer (15-15-15 N-P 2 O 5 -K 2 O) in order to ensure that all plants received sufficient nutrients. The pots were irrigated daily to field capacity until the plants had two leaves. Three plants per pot were selected and the rest removed. Five treatments of irrigation were established 24 days after sowing. The treatments differed in irrigation frequency and water volume applied and ranged from field capacity to extreme drought. In order to determinate the water volume necessary in each treatment, field capacity was measured in 6 pots only with soil, by soil weight before and after wetting and drainage resulting 0.217 cm 3 cm -3 . Also, in each irrigation event, the water volume applied was adjusted according to the pots weight which was incrementing by the plants biomass and it was calculated as field capacity percentage ( Table 1).
The experimental design was a factorial design with 8 repetitions. Factors established were species (two) and irrigation treatment (five) resulting 80 pots with three plants per pot. Additionally three pots per treatment, species, and time of determination (34, 48, 72, 88 DAS: days after sowing) were destructively sampled for biomass, resulting 120 pots. The total number of pots was 200 with three plants per pot.
Periodically, an exhaustive monitoring of the plants growing was performed. The variables evaluated per plant in each time of determination were: height by measuring leaves or panicles, total tiller number, total number of panicles, and distinguishing ripeness stage of panicles (emerged not flowering, beginning flowering, full flowering and beginning ripening and panicle ripe). Aerial biomass at the end of the assay was determined in 8 pots per treatment and species. Grain production was not evaluated due to dehiscence.
The results were analyzed using ANOVA test with the statistical package Statistix 8. Means were separated by application of the Tukey test. It was not necessary to transform the data because variances were homogenous.

Differences of traits between species and among treatments, and interaction between species and treatments
The ANOVA analysis showed that there were significant differences between the two species for all parameters measured (Table 2). Likewise, treatment within a species were significant (p < 0.001), except for maturing panicle percentage (p = 0.05). For most of the traits, significant interaction of species versus treatment was detected.

Growth and aerial biomass
The progression of height and aerial biomass shows that there were differences between species and responses to soil water availability (Fig. 1), both species being affected negatively by drought. Differences in height were very important within 34 days after sowing (DAS) in both species. In P. minor, the field capacity and light drought treatments differed greatly from those of severe and extreme drought for both traits. Moderate drought results were intermediate between the two extremes. However, in P. brachystachys this trend was different: severe and extreme drought regimens resulted in much lower heights and biomasses than those obtained at other water levels.
At 61 DAS differences in height between the two species were significant for light drought but not for other treatments (Fig. 2), although overall the response of both species in height was similar. However, biomass differences between species were significant for both field capacity and light drought treatments (Fig. 2),

Tillers and panicles
There were considerable differences between P. minor and P. brachystachys in terms of tiller and panicle numbers. Soil water availability had a strong influence (Fig. 3a-d) over these two traits. The greatest number of panicles were produced from 46 DAS to 61 DAS in P. minor (Fig. 3c) while in P. brachystachys this took place one to two weeks later (Fig. 3d).
P. minor produced significantly more tillers under field capacity and light to moderate drought levels than under severe and extreme drought conditions (Fig. 3a, 3c and 4). However, the only difference produced in panicle number was between moderate and severe/extreme drought levels (Fig. 4). In P. brachystachys, the maximum production of tillers and subsequently the maximum panicle number took place at the moderate drought level (Fig. 4).
The total number of tillers and panicles at f ield capacity were significantly (p < 0.01) smaller in P. brachystachys compared to P. minor. There were no other statistically important differences found for the rest of the treatments (Fig. 4).        Different letters indicate differences among the soil water conditions within P. minor (capital letters) and within P. brachystachys (lower-case). * p < 0.05, ** p < 0.01 and *** p < 0.001 represent the differences between species.

Panicle size
Panicle size for P. brachystachys was significantly lower than that of P. minor (p < 0.001) under field capacity, light, and moderate drought levels and p < 0.05 at severe and moderate drought levels (Fig. 5). Low water availability significantly reduced panicle size in both species. Differences in P. minor's panicle size were detected between the treatments with more water availability (field capacity and light drought) compared to moderate drought, as well as between these levels and those of severe and extreme drought (Fig. 5). The size of P. brachystachys' panicles was significantly different among field capacity, moderate, and extreme drought levels (Fig. 5).

Ripeness
Significant differences were observed between species in the percentage of panicles produced at 67 DAS according to ripeness ranks, although they were not found among treatments (Fig. 6). Between 82 to 91% of P. minor's panicles achieved ripeness levels (or exceeded anthesis stage), in contrast with 47 to 72% of those of P. brachystachys. There were no statistical differences among treatments for P. minor; however, P. brachystachys displayed an accelerated rate in the ripening process when water availability was reduced. This acceleration was manifested with significant statistical differences between field capacity and severe drought treatments (Fig. 6).

Discussion
This study showed that P. minor and P. brachystachys are signif icantly influenced in their growing by different levels of soil water availability. Although both species were affected in their development by drought, they exhibited a positive response to moderate levels of water stress. Development was, however, greatly affected by severe and extreme drought conditions. Plant growth was lower under drought stress, although differences between the two species were small. Both species produced seeds and overcame severe or extreme drought periods. This characteristic is typical of weed species in Mediterranean countries and allows them to remain in fields over long time periods. The distribution of both species in Europe (Tutin, 1980), and Southern Spain specifically (García-Baudín, 1983;Saavedra et al., 1989a;Jiménez-Hidalgo et al., 1997) agrees with this response to severe and extreme drought conditions. Field capacity conditions favoured P. minor over P. brachystachys that had higher biomass, tiller number, panicle number, and panicle length. These characteristics may allow P. minor to produce higher quantity of seeds and therefore more frequent infestations in irrigated fields of Mediterranean countries. This phenomenon has been observed previously by Saavedra et al. (1989a) in Southern Spain. This fact may also explain the Paleosubtropical distribution of P. minor that Pignatti (1982) observed, as well as the species' extension northwards into N.W. France (Tutin, 1980). Furthermore, P. brachystachys' biomass was not increased through exposure to field capacity or light drought levels in comparison with its response to the moderate drought regimen. On the contrary, its number of panicles was remarkably reduced under these conditions. Although at field capacity the panicles were, on average, longer than at other water levels, they also exhibited chlorosis, further indicating adaptation problems to excess of water in this species.
Irrigation in Guadalquivir river Valley produces temporary excess of water, hindering P. brachystachys, and therefore, favouring P. minor. Moreover, Om et al. (2004) showed that P. minor's seeds exhibited tolerance to anoxia during anaerobic respiration in rice. On the contrary, Ohadi et al. (2009) in germination studies found differences between irrigated and nonirrigated conditions in the P. minor seeds survival at 10 cm deep, with higher seed mortality under irrigated conditions, however, that did not happen at 20 and 40 cm deep.
P. brachystachys exhibited a higher tolerance to moderate water stress, producing more panicles without excessively affecting the mean size of panicle. Consequently, this species seems to be more adapted to dry land conditions than P. minor, which could explain the wider distribution of P. brachystachys across the Spanish provinces with a lower average rainfall (Saavedra et al., 1989a).
Under the experiment conditions, P. minor showed a faster growth rate, in accordance with Afentouli and Eleftherohorinos (1999). In addition, the phenological cycle of P. minor was shortened compared to P. brachystachys and appeared not to be dependent on the level of water availability. However, P. brachystachys had the capacity to reduce the time of ripeness for panicles under water stress conditions. This may be an adaptive strategy to dry conditions in the Mediterranean climate, where the rainfall is scarce and irregular.
The results obtained show that the growth and reproductive traits such as biomass, tiller number, or panicle number in both species are influenced by the water availability. This different behaviour of P. minor and P. brachystachys depending on the soil water levels during the plant development may complicate the assessment of the competitive capabilities if those parameters are used. In our experiment under moderate stress treatments P. brachystachys developed a greater number of panicles regarding others treatment. In contrast, P. minor produced more panicles when water levels were high (field capacity) which agrees with results obtained by Afentouli and Eleftherohorinos (1999). This is an important consideration because it can limit the use of competition models based solely on panicle numbers, such as that proposed by Jiménez-Hidalgo et al. (1997), or biomass. Therefore, the competitive studies based on models which use panicle number or biomass to explain the competitive relationship between P. minor or P. brachystachys and cereal crops, should take into account the conditions of water availability under which they are performed.
The results of this study, though it was performed once and in controlled conditions, are totally reliable by the great number of repetitions (300 plants of each species) and the exhaustive monitoring of variables evaluated per plant such us height, biomass, tiller number, panicle number and length and different ripeness stage of panicles. This great number of variables could not have been evaluated in field conditions with this detail nor the different treatments of water availability established. However, field trials with lesser number of variables and treatments are mandatory in medium or short-term in order to validate the results obtained in this experiment.
As final conclusion, this study has revealed differences in the responses to different soil water levels for P. minor and P. brachystachys. Both survived even when were exposed to extreme drought, although, both were negatively influenced in their development by severe and extreme drought levels. Field capacity conditions consistently favoured P. minor and harmed P. brachystachys. The observed tolerance of P. brachystachys to moderate drought (as shown by a greater production of tiller number, panicle number, and an accelerate ripeness), together with the negative effects caused in plants by the field capacity treatment, may explain its wider distribution in rainfall fields and its adaptation to moderate drought conditions. Their distinct response to different soil water levels could explain differences in chorology and ecological preferences between these two species.