Environmental and Experimental Botany 55 (2006) 61–69 Characterizing genotype specific differences in survival, growth, and reproduction for field grown, rapid cycling Brassica rapa Martin G. Kelly∗ Department of Biology, Buffalo State College (SUNY), 1300 Elmwood Avenue, Buffalo, NY 14222-1095, USA Accepted 30 September 2004 Abstract Rapid cycling Brassica rapa (RCBr) develops rapidly, and has both small adult size and a brief life cycle. Yet, in spite of many investigations using RCBr, extremely few plant ecologists have used this plant in the field. This study is the first to describe the genotype specific variation in traits describing survival, growth, and reproduction for field grown, RCBr. I also identify traits associated with fitness. Five genotypes of RCBr were used: standard, anthocyaninless, yellow-green, anthocyaninless and hairless, and anthocyaninless and yellow-green. Plants were grown outside in a “common garden”. Eight plant traits were measured: life span, height, growth rate, leaf size, number of flowers and fruits, fruit set, and fitness. All traits, except life span, differed significantly among the five plant genotypes. Correlation analysis revealed that fitness increased as each of these of seven plant traits increased. This study demonstrates that RCBr can serve as a model organism in ecological field studies. © 2004 Elsevier B.V. All rights reserved. Keywords: Field ecology; Rapid cycling Brassica rapa; RCBr; Model organism 1. Introduction B. rapa has established weedy and naturalized popula- tions in North and South America, Australia, and Asia One species that botanists have broadly employed in (Warwick and Francis, 1994, Part V,p. 19). Thus, many all areas of research is Brassica rapa (syn. campestris). field studies have been done with it to understand the This species is endemic throughout Europe eastward consequences of field release of genetically modified to Siberia (Warwick and Francis, 1994, Part V, pp. organisms (Jorgensen and Andersen, 1994; Hauser et 11, 19). This plant species is also widely cultivated in al., 1997, 1998a, 1998b; Snow and Palma, 1997; Snow cooler climates (Pak Choi, Turnip Rape, Choy Sum, et al., 1999; Pertl et al., 2002; Halfhill et al., 2003; Zhu Chinese Cabbage, Tendergreen, Turnip, Sarson, and et al., 2004). Broccoli Raab; Williams and Hill, 1986). Moreover, Of special importance to this investigation is rapid cycling B. rapa (RCBr). RCBr was derived using classical methods of artificial selection and breeding ∗ Tel.: +1 716 878 4608; fax: +1 716 878 4028. (Williams and Hill, 1986). Under optimal conditions E-mail address: [email protected]. 0098-8472/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.envexpbot.2004.09.012 62 M.G. Kelly / Environmental and Experimental Botany 55 (2006) 61–69 Table 1 rapa grown in growth chambers and the field concluded Selected ecological studies conducted with Brassica rapa presented that it was not possible to use the phenotypic outcomes in chronological order for plants grown in growth chambers as predictors for References Brassica rapa Experiments in plants grown in the field. Miller and Schemske (1990) Rapid cycling Growth chamber Though several ecological investigations have Agren and Schemske (1992) Rapid cycling Growth chamber used RCBr, ecologists—except Kelly and Terrana Agren and Schemske (1993) Rapid cycling Growth chamber (2004)—have not employed RCBr in the field (Table1). Agren and Schemske (1994) Naturalized Field and greenhouse Before RCBr can be applied as a model organism to Jorgensen and Andersen Naturalized Field more substantial ecological questions two points must (1994) be established. First, that ecologically relevant pheno- Miller (1995) Rapid cycling Greenhouse typic variation exists in field grown plants. Second, that Nakamura et al. (1995) Naturalized Field rapid cycling forms of this species grown in a natural Schmitt et al. (1995) Rapid cycling Greenhouse Davis et al. (1996) Rapid cycling Growth chamber setting can be used as model for native or naturalized B. Gurevitch et al. (1996) Rapid cycling Greenhouse rapa. Here, I characterize the growth of RCBr in a field Klaper et al. (1996) Rapid cycling Greenhouse setting. I identify some individual traits associated with Miller (1996) Naturalized Greenhouse plant survival and reproduction in a field grown popu- Mitchell-Olds (1996) Naturalized Greenhouse and lation of RCBr. I see this fundamental goal only as a growth chamber Pilson (1996) Naturalized Field first step. Ultimately, the second goal must be explicitly Hauser et al. (1997) Naturalized Field met before ecologists adopt RCBr as a model organism Hauser et al. (1998b) Naturalized Field for native or naturalized B. rapa in experiment based, Hauser et al. (1998b) Naturalized Field field research. Siemens and Mitchell-Olds Naturalized Field and growth (1998) chamber Stowe (1998) Rapid cycling Laboratory Kleier et al. (1999) Rapid cycling Growth chamber 2. Materials and methods Sleeman and Dudley (2001) Rapid cycling Greenhouse Pertl et al. (2002) Naturalized Field 2.1. Plant material Siemens et al. (2002) Naturalized Field Hauser et al. (2003) Cultivar Field Kelly and Terrana (2004) Rapid cycling Field RCBr was derived from a global collection of B. rapa (L.) varieties (Williams and Hill, 1986). Plants were selected for the following six qualities: reduced RCBr has a life cycle of 35–40 days, from parental size at maturity, minimum time from germination to seed sown to offspring seed harvest (Williams and Hill, flowering, uniformity of age at first flowering, high 1986). Compared to normal B. rapa, which can produce flower production, rapid maturation of seeds, and lack two generations in a year, under optimal conditions of seed dormancy (Tomkins and Williams, 1990). In- RCBr can produce 10 generations in a year. Thus, the dividuals that flowered fastest were used as the base potential applications of RCBr to experimental botany population. These individuals were out-crossed to gen- are diverse (Musgrave, 2000). erate seeds. In the next generation, the 10% of the off- After Williams and Hill (1986) summarized the spring population that flowered first were selected as results of their selection for rapid cycling Bras- parents. These 288, or more, plants were mass polli- sica species, plant ecologists were among the early nated to produce the next generation of seeds; artifi- adopters. Ecologists have since employed both nat- cial selection continued until the response to selection uralized and rapid cycling varieties of B. rapa in a was stabilized (Williams and Hill, 1986). Under op- wide variety of studies (Table 1). Though Miller and timal laboratory conditions, RCBr flowers within 16 Schemske (1990) were the first to publish ecological re- days of seed germination (Williams and Hill, 1986). It search using RCBr, almost all of our knowledge about is important to recognize that RCBr retains consider- the growth and development of RCBr is based on plants able isozymic variation (Williams and Hill, 1986). In raised in controlled environments (Table 1). However, addition, when inbreeding is forced, fitness is signifi- Torresen and Lotz (2000) in a direct comparison of B. cantly reduced (Evans, 1991). Both outcomes suggest M.G. Kelly / Environmental and Experimental Botany 55 (2006) 61–69 63 Table 2 Five genetically defined varieties of RCBr known to differ in phenotype Variety Genotype Phenotype Standard Rbr/Rbr Normal chlorophyll production, anthocyanin expression, and epidermal hair density Anthocyaninless anl/anl Normal chlorophyll production and epidermal hair density, lacks anthocyanin Yellow-green ygr/ygr Deficient in chlorophyll production, normal anthocyanin expression and epider- mal hair density Anthocyaninless and hairless anl/anl and Hir (0–1) Normal chlorophyll production, lacks anthocyanin and has a very reduced num- ber of epidermal hairs Anthocyaninless and yellow-green anl/anl and ygr/ygr Deficient in chlorophyll production and lacks anthocyanin, has normal epidermal hair density that B. rapa did not pass through a genetic bottleneck mental differences, seeds were randomly assigned to in order to establish the rapid cycling lines. one of the five plots and to a grid position in the plot. RCBr are available in a wide variety of known geno- Seed sowing positions were marked as a grid with a types with distinct phenotypes. For this experiment, I #804 SoilMasterTM dibble board (6 × 8 pegs spaced compared five self-incompatible genotypes known to 4.4 cm× 6.4 cm apart). Each 6 × 8 plot of 48 seeds was differ in plant pigmentation and trichome production 53.5 cm wide and 27 cm tall. All seeds were sown on (Table 2). I chose phenotypic traits that alone, or in August 29, 2002. combination, might affect plant survival or reproduc- Experimenter sown seeds were used for two reasons: tion. I used 48 seeds per genotype, for a total of 240 (1) the randomization of individuals across the envi- seeds. I used seeds purchased from Carolina Biolog- ronment minimizes the correlation between genotype ical Supply Company (USA). RCBr seed stocks for and environment, or the correlation of the phenotype these same five self-incompatible genotypes may also with the environment because of past environmental ef- be purchased from Blades Biological Ltd. (UK). fects (Mitchell-Olds and Shaw, 1987) and (2) Wade and Kalisz (1990) convincingly argued that the application 2.2. Field study design of quantitative methods to measure natural selection was not sufficient to determine why selection operated Plants were grown outside in a “common garden” in the manner observed; they recommended the use experiment and experienced the same general grow- of experimental manipulation to identify the agents of ing conditions. Over the course of the experiment viability and fecundity selection as a complement to mean daily high and low air temperatures were 24.1 quantitative analysis. and 14.6 ◦C, and mean daily precipitation was 0.2 cm. Six students in a Plant Ecology course collected Weather data were recorded at a station located 2.3 km data. This semester long course met one evening per from the field site.