Seed-related traits and their adaptive role in population differentiation in Avena sterilis along an aridity gradient

Four populations of the annual grass A. sterilis distributed along an aridity gradient in Israel (Mount Hermon, Northern Galilee, Shefela and the Negev Desert), were studied to: 1) reveal a general pattern of seed dormancy and persistence in the soil
of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  © 2009 Science From Israel / LPPLtd., Jerusalem  Israel Journal of Plant Sciences Vol. 57 2009 pp. 79–90 DOI: 10.1560/IJPS.57.1–2.79 E-mail: Seed-related traits and their adaptive role in population differentiation in  Avena sterilis  along an aridity gradient S ERGEI  V OLIS  Department of Life Sciences, Ben-Gurion University of the Negev, P.O. Box 653, Be’er Sheva 84105, Israel (Received 8 July 2008; accepted in revised form 26 November 2008) ABSTRACT Four populations of the annual grass  Avena sterilis  distributed along an aridity gra-dient in Israel (Mount Hermon, Northern Galilee, Shefela, and the Negev Desert), were studied to (1) reveal a general pattern of seed dormancy and persistence in the soil seed bank in this species; (2) compare seed size and demography of reciprocally introduced seeds and seedlings; and (3) test the adaptive nature of the observed patterns. The steep aridity gradient in Israel represents two parallel clines of envi-ronmental productivity (annual rainfall) and predictability (variation in amount and timing of annual rainfall). The four populations examined represented the following environments: (1) desert—low productivity and predictability, drought stress; (2) semi-steppe batha—moderate productivity and predictability; (3) grassland—high productivity and predictability; and (4) mountain—high productivity and predict-ability but with severe frost stress. The highest proportion of dormant seeds, most sequential germination of the rst and the second orets of a spikelet over three years, and highest importance of desiccation tolerance were found at the desert location, consistent with bet-hedging buffering against unpredictability of rainfall and high probability of drought in this environment. Signicant population srcin by environment interactions were observed for yield and reproductive biomass, but no advantage of local ecotype was detected for these two traits. However, another tness component, seedling survival, showed not only the interactive effect of srcin and locality, but also the superiority of the local ecotype and decreasing tness rank from indigenous ecotype towards the most environmentally dissimilar ecotype, sug-gesting local ecotype adaptation of seedlings. There was a genetically determined decrease in seed mass with increase in aridity without concomitant effect of frost. The selective forces that may differentially affect seed size along the aridity gradient are competition, predation intensity, importance of dispersal distance, and bet-hedg-ing against rainfall unpredictability. Further experiments are needed to determine the precise nature of aridity-related evolution of seed size in  A. sterilis . Keywords : ecotype, local adaptation, plant strategy, reciprocal transplant, wild oat INTRODUCTION Population differentiation through selection processes is a well-documented phenomenon occurring in a locally specic environment that imposes selection intensity sufcient to override gene ow (reviewed by Bradshaw, 1965; Linhart and Grant, 1996). Since the classical stud-ies of Turesson (1922) and Clausen et al. (1940, 1948), cross-environment comparisons in plants have been used to reveal the adaptive nature of inter-population phenotypic variation. Besides the goal of detecting local adaptations, the emphasis of these studies was either on phenotypic traits that evolve in particular environments (e.g., Gauthier et al., 1998; Leiss and Muller-Scharer, 2001; Housman et al., 2002) or environmental factors responsible for ecotypic differentiation (e.g., Kindell This paper has been contributed in honor of Azaria Alon on the occasion of his 90th birthday.   Israel Journal of Plant Sciences 57 2009 80et al., 1996; Cheplick and White, 2002). Survival at different stages of the life cycle and fecundity are the major estimates of plant tness, and in most studies of local adaptation these parameters were of major con-cern. However, there is a growing number of studies where reciprocal introductions are applied to investigate adaptive inter-population variation in different morpho- logical and life history traits, including phenology (Ben -nington and McGraw, 1995), leaf and root morphology (Gurevitch, 1992; Kerley, 2000; Housman et al., 2002), leaf demography (Lovett Doust, 1981), vegetative reproduction (Fritsche and Kaltz, 2000), growth and resource allocation (Chapin and Chapin, 1981; Leiss and Muller-Scharer, 2001), and importance of the seed bank (Philippi, 1993; Cavieres and Arroyo, 2001; Volis et al., 2004).Adaptive variation in traits, such as seed size and dor-mancy, has been a subject of much research (reviewed in Howe and Swallowed, 1982; Ellner, 1987; Rees, 1996). Despite a small scale of spatial environmental variation that a seed usually experiences, and a large maternal effect that may mask genetic variation of the offspring, heritable differences in seed size and dormancy are expected to evolve in environments that differ in tem-poral heterogeneity, i.e., predictability of environment (Cohen, 1966; Venable and Brown, 1988).In conspecic plant populations distributed along clear environmental gradients, e.g., soil nutrient and water gradients, life history, morphological and physi-ological traits may change in a predictable way as a re-sponse to major environmental factors if they are under selection. If temporal heterogeneity is associated with the major gradient, seed dormancy may constitute a part of such adaptive response and predictably change along a gradient (Cavieres and Arroyo, 2001). Other factors, such as spatial heterogeneity and intensity of predation, may also correlate with the major gradient, and affect seed size (Howe and Swallowed, 1982; Reader, 1993; Price and Joyner, 1997).In this study, I used reciprocal seed and seedling introductions to compare the germination over three years, and several life history parameters, including seed size, in plants of an annual grass,  Avena sterilis , from four populations distributed along an aridity gra-dient. In Israel the north–south aridity gradient creates steep climatic and ecological clines over relatively short distances (Bitan and Rubin, 1991; Aronson et al., 1992; Kadmon and Danin, 1997). Water is the main limiting and uctuating resource in this area that creates a severe productivity–predictability gradient from mesic Medi-terranean to xeric desert (Aronson et al., 1992). The ob- jectives of this study were the following: (1) to reveal a general pattern of seed dormancy and persistence in the soil seed bank in this species; (2) to compare seed size and demography of reciprocally introduced seeds and seedlings; (3) to test the adaptive nature of differences between the four studied populations. MATERIALS AND METHODSStudy species  Avena sterilis  L. (Poaceae) is a winter annual, and a predominantly selng grass. This species is one of the major components of the annual vegetation throughout Israel. It forms massive stands in the open park-forests with Quercus ithaburensis , aggressively colonizes op-ened-up maquis and hemicryptophytic/dwarf shrub formations, and penetrates into favorable desert habitats (wadi beds and loessy depressions) (Zohary, 1983).This species is synaptospermous, i.e., the whole spikelet disarticulates and acts as a drill-type dissemi- nation device. Each spikelet has two to four orets, of which the most basal or primary oret and one above, the secondary, are usually functional, while the remain-ing, if any, are often undeveloped and nonfunctional. The inorescence is a panicle with basipetal maturation of spikelels (from the periphery to the center). Within the spikelets, however, orets mature acropetally, from the basal oret up. Seeds are produced in spring (April– May). Seedlings emerge in November–December, grow and mature through winter–early spring, and senesce before summer. Seeds that do not germinate in the autumn following dispersal either die or enter the soil seedbank where they can remain dormant for several years (Sanchez Del Arco et al., 1995). Choice of populations Seed collections of  A .  sterilis  were made in 1996 from 20 locations, employing nested sampling design, i.e., in four regions with ve populations per region. Each cluster of ve populations was representative of one of the following environments/vegetative communities: desert, semi-steppe batha (open Mediterranean vegeta-tive community dominated by sub-shrubs), Mediterra-nean grassland, and mountain. Populations comprising a cluster were ≤5 km (mountain) or ≤20 km (other three groups) from each other and represented the same environment with respect to relief, slope exposition, vegetation, and soil type. All 20 populations were used in a study of population genetic structure (Volis et al., unpublished), and one population from each cluster was used for a comparative study of plant life histories and test of local adaptation. The locations of populations representing each habitat type are the same as those described in Volis et al. (2002b) and I kept the same abbreviations in this paper:  Volis / Population differentiation in Avena sterilis 81 (1) SB location. Lat. 30°51 ¢ N, Long. 34°46 ¢ E. Wadi in the Negev Desert, 3 km southwest of Sede Boqer fenced as an experimental area of the Mitrani Department for Desert Ecology. (2) BG location. Lat. 31°36 ¢ N, Long. 34°53 ¢ E. North- facing slope in the Shefela Hills in the Beit Guvrin National Park. (3) AM location. Lat. 32°55 ¢ N, Long. 35°32 ¢ E. North-facing slope in the Upper Galilee, 1 km west of Kibbutz Ammiad.(4) MH location. South-facing slope on Mount Hermon (Hermon Nature Reserve).The three locations for which rainfall data are avail-able fall along a south-to-north gradient with increasing amounts and predictability of precipitation (Fig. 1), as reected in long-term observation data (90, 408, and 580 mm of average annual rainfall; 43, 32, and 29% coefcient of variation in annual rainfall over 42, 49, and 49 years of observations, respectively). The des-ert population is located in the wadi and therefore in favorable years may receive amounts of water close to that in more mesic locations due to water accumula-tion from runoff from the adjoining hills. However, in unproductive years, when low amounts of rainfall are coupled with unfavorable distribution of rainy events, extremely high mortality of seedlings/young plants or even no germination of  A. sterilis  seeds is observed. The mountain population (MH) is located at the elevation of 1,500 m and it receives the highest amount of rainfall (>1,300 mm, which is the rainfall amount at Majdal Shams, a few kilometers away, Israel Meteorological Service data for 1969–1972). The climate at MH is much cooler than in Ammiad (the area is covered with snow during winter months) and is very predictable in both rainfall and frost occurrence (no reliable data avail-able, pers. observ.). Thus, in addition to the clinal effect of aridity present in all environments, the mountain population also experiences a discrete environmental effect of frost absent in other locations. EXPERIMENTAL DESIGNSeed introduction experiment In 1998/99 around 200 spikelets collected from green-house-grown mother plants from each of the four popu-lations were buried at each transplant site. A spikelet was placed in a separated cell of a plastic tray lled with soil of the transplant environment and covered with ne Fig. 1. A. The four populations (= introduction sites) and a scheme of experimental design. B. The annual rainfall and its year- to-year variation (CV) at 25 meteorological stations of Israel over more than 40 years of observations. C. The annual rainfall and its CV in seasons 1997/98 and 1998/99 at the introduction sites. The meteorological stations closest to the populations SB, BG, and AM (all within 3 km distance) are designated.   Israel Journal of Plant Sciences 57 2009 82metal net to prevent seed predation by ants, rodents, and birds. In Mediterranean winter annuals, mass germina- tion occurs after the rst effective rain (>10 mm of rain - fall); therefore, two months after the rst effective rains, I removed the trays, brought them to the laboratory, and classied the seeds into the following groups: (i) germi -nated (with the radicle protruded) but desiccated before coleoptile development and able to develop adventitious roots upon wetting (as determined in separate tests); (ii) germinated and developed coleoptile or true leaves; (iii) non-germinated. In 1999/2000 the non-germinated spikelets were buried again at the respective transplant sites and the procedure was repeated; this was repeated again in 2000/2001. The two-month-old plants devel-oped from introduced seeds were dried at 70 ºC and weighed. Seedling transplant experiment A reciprocal transplant experiment was conducted in four locations during two consecutive years (1997/98 and 1998/99). Unfortunately, for the two extreme envi- ronments, MH and SB, I only have data from one year. The mountain plot (MH) was destroyed in 1997–1998 by grazing cows, and the desert plot (SB) received too little rain in 1998–1999 to let oat plants mature and pro-duce seeds. At each site a 50–100 m 2  plot was fenced and cleared of vegetation either at the site where the original population was collected (BG and MH) or in close prox - imity (500 m away at SB; 3 km away at AM). I used the self-progeny of the ten mother plants/site grown under uniform greenhouse conditions in win- ter 1996. This was done to remove the site-specic seed maternal effect. The plants were placed singly in 10-liter buckets lled with a potting mixture of equal parts peat and vermiculite, spaced 1 m apart. Irriga-tion was provided 3 times a week through 1-liter/hour drippers. Upon harvesting, the seeds were pooled and transplanted in each environment within two weeks after the rst effective rain (>10 mm). A randomized block design was established with each block containing two-week-old seedlings from all four populations arranged as a lattice pattern of 4 by 4 plants, with 10-cm spacing. At seed maturity, I calculated the number of panicles per plant and number of spikelets in each panicle. From this I estimated the yield (total number of spikelets) per plant. Around 100–200 spikelets were collected from each pop-ulation/block and used to estimate the average spikelet weight and reproductive biomass per population/block. Data analysis I used a 3-way ANOVA to determine the possible inter-actions between srcin of seeds, site of transplanting, and year (all xed effects), and a 2-way ANOVA to ana - lyze the srcin by transplant site interaction (both xed effects). The former was done for BG and AM sites only and the latter for all four sites. Block nested within site was added as a random effect to both models. Data required no transformation to satisfy the assumptions of normality and homoscedasticity. The Tukey–Kramer test was used for population comparisons across sites.The differences in seed germination and survival of seedlings among populations within introduction sites were determined by a χ 2  test. The association of site and srcin was analyzed by a test of independence.Population responses across sites were compared by Tukey–Kramer test (reproductive traits and weight of two-month-old plants) and χ 2  test (seed survival). In the latter test the expected frequencies for all introduction sites were assumed not to differ from frequency at the site of srcin. RESULTSSeed introduction experiment The effect of introduction site on spikelet germination was most important. In the rst year after introduction, the lowest percentage of spikelets germinated (with either two orets germinated, one oret germinated, or protruded radicle that subsequently desiccated) was observed at the SB site (range 38.4–47.3%), was high at the BG and AM sites (range 81.0–98.2% and 82.6–92.0%, respectively), and intermediate at the MH site (range 73.5–98.1%). However, the effect of plant srcin was evident at the MH site, where germination percentage was signicantly higher for spikelets of MH srcin in the rst year after introduction (98.1 vs. 78.5, 73.5, 77.5 for AM, BG, and SB, respectively, χ 23  = 24.5,  p  < 0.001) and over three years (98.1 vs. 81.0, 75.3, and 82.3 for AM, BG, and SB, respectively, χ 23  = 21.5,  p  < 0.001). In the second year after introduction, the number of spikelets germinated from those dormant in the previ- ous season was high at the SB site (range 8.5–40.9%), much smaller at the BG and AM sites (range 0.4–6.6% and 0–6.1%, respectively), and negligible at the MH site (range 0–1.8%). At two sites, BG and AM, germination percentage in the second year after introduction was higher for spikelets of SB srcin ( χ 23  = 6.1 and 7.2,  p  < 0.05). In the third year after introduction, spikelets dor- mant in two previous seasons germinated only at SB and MH sites (range 2.3  –  5.9% and 0–4.3%, respectively). At the MH site, germination percentage in the third year after introduction was higher for spikelets of SB srcin ( χ 23  = 8.1,  p  < 0.05).Spread of germination over time was more pro- nounced for spikelets of SB srcin at the SB site than  Volis / Population differentiation in Avena sterilis 83for other ecotypes when per-year germination fraction was calculated from the total number of seeds germi- nated (Table 1). The highest proportion of seeds of SB srcin germinated in the second year, while in the other three ecotypes the rst year germination predominated (Table 1).Survival of seedlings from spikelets that germinated (i.e., having roots and coleoptiles or true leaves) was affected by both introduction site and srcin, as well as their interactive effect (Table 1). At the BG and AM sites, seedling survival was uniformly high (range 83.1– 100%), while at the SB and MH sites seedling survival dramatically decreased in opposite directions from SB to MH plants (Table 1).The association of site and srcin, analyzed by a test of independence, was not signicant for germination fraction in the rst year after introduction ( χ 29  = 5.4,  p  > 0.05), but was highly signicant for seedling survival in the rst year after introduction ( χ 29  = 148.5,  p  < 0.001).Seed viability of  A. sterilis  in the soil was found to not exceed 3 years and the pattern of oret germination to correspond to that of maturation, viz. the basal oret rst, then the one above and so on. However, germina - tions of third and fourth orets were rare events (less than 1% of germinated spikelets) and were not account- ed for. Germination of two orets in the same spikelet in different years (viz. germination of the second oret a year after the basal oret germinated) was estimated in this experiment for the SB site only, so it could not be compared for the four ecotypes across environments. Besides, as seedlings were dug up at the age of two months, their aboveground part was cut off and only the belowground part containing the non-germinated second oret was buried again, so we should be cautious with interpretation of observed germination pattern in the same spikelet over time as naturally occurring. The Table 1Germination fractions and survival of germinated spikelets recorded two months after germination (in %) after reciprocal seed introduction, and results of χ 2  test on effect of srcin for each site. Germination fraction was estimated during three years and survival of seedlings was estimated for the rst year after introduction (in the second and third years, the sample size was too small for reliable estimation of survival). Spikelet was considered germinated if at least one oret germinated. Germination fraction is calculated as either percentage of buried spikelets or percentage of germinated spikeletsSite/population Germination fraction Germination fraction Survival of germinated (% of buried spikelets) (% of germinated spikelets) spikelets (%) 1998/1999 1999/2000 2000/2001 1998/1999 1999/2000 2000/2001 1998/1999 SB site SB 9.7 18.9 5.9 28.12 54.78 17.1 100BG 21.3 7.3 2.4 68.71 23.55 7.742 86 AM 21.1 19.7 2.3 48.96 45.71 5.336 13MH 30 16.4 5.9 57.36 31.36 11.28 6 χ 2  12.7** 7.4 ns 3.2 ns 35.3*** 24.9*** 8.4* 282.8*** BG site SB 78.6 6.6 0 92.25 7.746 0 83BG 91.5 3 0 96.83 3.175 0 92 AM 92 3 0 96.84 3.158 0 91MH 98.2 3 0 97.04 2.964 0 100 χ 2  22.7*** 2.6 ns 4 ns 4 ns 18.5*** AM site SB 85.7 6.1 0 93.36 6.645 0 100BG 84.3 2 0 97.68 2.317 0 99 AM 92 3 0 96.84 3.158 0 100MH 82.6 0 0 100 0 0 100 χ 2  4.2 ns 7.2 ns 7.7 ns 7.4 ns 1.7 ns MH site SB 77.5 0.5 4.3 94.17 0.608 5.225 86BG 73.5 1.8 0 97.61 2.39 0 93 AM 78.5 0.5 2 96.91 0.617 2.469 95MH 98.1 0 0 100 0 0 96 χ  2  24.5*** 2.6 ns 8.1* 6.3 ns 3.6 ns 9.9* 8.2* Level of signicance: ***  p  < 0.001, **  p  < 0.01, *  p  < 0.05, ns—not signicant.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks