Impacts of greenhouse fertilization and planting season on survival and reproductive potential of Silene regia transplants

Restoration of rare plants is essential to maintain species diversity in natural areas. The methods used to grow and plant these rare plants for restoration impact seedling health and potentially the success of restoration attempts. We evaluated how different fertilizers and planting seasons impacted survival and reproductive potential for transplants of Silene regia. Plants were grown from seeds in greenhouses with four fertilizer treatments: slow‐release, liquid biweekly (every other week), liquid weekly, and control (no fertilizer). Transplanting was conducted in spring and fall for 2 years (2010 and 2011) at three sites. Number of leaves on transplants were counted before planting. Plants were counted over the growing seasons for 8 years (2010–2017) to determine survival and stem elongation (an indication of reproductive potential). Transplants produced the most leaves with slow‐release, followed by both liquid fertilizers and the fewest leaves with no fertilizer. The fertilizer type had a limited effect, but plants grown with fertilizer had increased survival and stem elongation relative to no fertilizer when a fertilizer effect was found. Planting season had a more consistent effect, with spring having higher survival and stem elongation. Sites differed for all measured parameters. Year‐to‐year survival was initially high, then dropped after 3–5 years, and less than 20% of plants remained by 2017. These findings demonstrate the value of pre‐planting fertilizer to provide vigorous transplants and planting season for reintroductions. Finally, multiple sites and long‐term monitoring are important to ensure success for reintroductions of rare species such as S. regia.


Introduction
Silene regia Sims (Caryophyllaceae), commonly called royal catchfly, is a long-lived forb with brilliant red flowers pollinated primarily by hummingbirds (Menges 1995;Fig. 1), found in wet-mesic to dry habitats of prairies, woodlands, and glades (Dolan 1994).This species was known historically from 13 states in the Midwest and southeastern United States (AL, AR, FL, GA, IL, IN, KS, KY, MO, MS, OH, OK, TN) (NatureServe 2022).Habitat loss, woody encroachment, and poaching have left this species rare throughout its range.S. regia is listed as globally vulnerable and as state critically imperiled in six states, imperiled in three states, and possibly extirpated from two states (NatureServe 2022).In Illinois, S. regia is critically imperiled and reported in 22 counties, including many reintroductions (Illinois Department of Natural Resources [IDNR] 2022).For surveys where plants were found in Illinois since 2015, populations range in size from 1 to 278 individuals, but 83% had fewer than 50 individuals (IDNR 2022).
Small populations are at greater risk of extinction as a result of stochastic events or their inability to adapt to environmental changes due to reduced genetic diversity.Populations at the edge of their range are often smaller and have decreased genetic diversity and increased genetic differentiation (Eckstein et al. 2006;Eckert et al. 2008).These small populations also may have reduced reproductive success (Reed 2005).In the eastern part of its range, small populations of S. regia have shown many of these issues.Dolan (1994) found that populations in Indiana and Ohio had lower genetic variation compared to western populations, and genetic variation was correlated positively with population size.Smaller populations of S. regia also had reduced germination percentages, possibly as a result of lower genetic diversity (Menges 1991).The appearance of S. regia seedlings is episodic, and they often have very low survival (Menges 1995;Menges & Dolan 1998).Recruitment was enhanced by fire, and without this recruitment, some populations of S. regia are expected to go extinct in Midwest prairies (Menges & Dolan 1998).Restoration and intervention may be necessary to secure S. regia's persistence throughout its range (Kull 2020).
Introductions of plants to new sites is one approach to ensure the survival of rare species and may be beneficial for S. regia.A study by Menges (1988) found S. regia was not limited in its distribution only due to habitat requirements.S. regia also is found in approximately 17-33% of remnant prairies in eastern North America, suggesting it may have been more common on the landscape than it is currently (Menges 1991).Seedling recruitment is believed to be limited by a lack of pollination, fruit and seed predation, and fire suppression (Edgin & Mankowski 2013).Many restoration attempts have been made for S. regia across Illinois (Edgin & Mankowski 2013).Several attempts in southeastern Illinois using transplants between 1993 and 2007 had limited success (Edgin 2012).
Unknowns exist on how best to achieve successful introductions.Menges (2008) stated how ecologists define introduction success (e.g.population reproduction), but reports on the outcomes are lacking (Godefroid et al. 2011).Often it is only successful introductions that are reported; however, a lack of long-term monitoring may be inflating the numbers of truly successful introductions (Menges 2008;Godefroid et al. 2011).Godefroid et al. (2011) found that monitoring usually ceased after 4 years, which may not be long enough to capture recruitment or reversals in initially high survival estimates of long-lived perennials.Adding to the complexity, many factors influence reintroduction success, including introduction techniques (e.g.fertilization; Guerrant & Kaye 2007;Kaye 2009).
Fertilization of seedlings is one technique that has the potential to alter the survival of introduced plants.Fertilizers are used to produce robust seedling growth that will hopefully translate to higher survival after transplanting.Plant nurseries commonly use an immediately available, water-soluble fertilizer (Jacobs & Landis 2009).This approach is often economical and allows for a quick and uniform liquid application of the fertilizer with watering, but these nutrients are not available after the introduction of plants into natural areas.Another approach is the use of a slowrelease fertilizer mixed into the potting media.These slow-release fertilizers are often more expensive, but offer extended nutrient supply potentially even after the plants are introduced (Jacobs & Landis 2009).The effect these slow-release fertilizers have on plant species growth is often species-specific.Few recommendations are found for native species (Jacobs & Landis 2009) and fertilizers may not always be beneficial for species introductions.
In the case of seedlings of Penstemon tubaeflorus (white wand beardtongue) grown in growth chambers, Annis et al. (2014) reported increased shoot mass, number of leaves, and root length in seedlings with slow-release fertilizer and no increased growth with liquid fertilizer compared to control with no fertilizer.For Schizachyrium sanguineum (crimson bluestem), controlled release fertilizer resulted in increased growth of seedlings in a plant nursery (de Oliveira et al. 2022).These positive effects of fertilizers on seedling growth during production may or may not parallel the survival of seedlings after transplanting in the field.Larger transplants of three sedge species (Carex lanuginosa-woolly sedge, Carex nebrascensis-Nebraska sedge, and Carex rostrata-beaked sedge) had greater survival than smaller transplants in riparian meadows (Steed & DeWald 2003).However, studies reporting how fertilizer additions at the time of planting for restorations showed variable effects on the survival of native species and sometimes were dependent on planting season (Kaye 2004;Kaye & Brandt 2005;Herriman et al. 2016).Studies by Kaye and Brandt (2005) also reported various impacts of fertilizer at the time of planting on flowering for the restoration of native species.All these studies show that fertilizer effects on the survival and flowering of native transplants were variable and species-dependent.No studies reported how fertilizer during the production of transplants affected plant survival and flowering for in situ restorations.Along with fertilization, the season when seedlings are planted can play a large role in the success of an introduction, as environmental factors (e.g.soil moisture, temperature) important to the growth and survival of plants vary by season (Copeland et al. 2022).Similar to fertilization, responses to planting season are species-dependent.When comparing spring vs. fall plantings, spring plantings for restoration increased the survival of transplants for some native species (Kaye & Brandt 2005;Page & Bork 2005;Copeland et al. 2022); while fall plantings increased survival for other species (Kaye & Brandt 2005;Page & Bork 2005).How planting season affected flowering was also reported by Kaye and Brandt (2005) where flowering was greater when planting in spring than in fall.Planting season was one of the most important factors for successful groundcover restoration in forests based on compiled data for a variety of species (Trusty & Ober 2011).
This study evaluated how fertilization (liquid, slow-release) affected the seedling growth of S. regia transplants in a greenhouse.In addition, this study evaluated how these fertilization treatments, along with planting season (spring and fall) in 2 years (2010 and 2011) affected survival and reproductive potential (stem elongation) of S. regia over seven and eight (2010-2017) years at three sites in the Prairie Ridge State Natural Area (Illinois, U.S.A.).Understanding how different fertilizer treatments and planting seasons affect S. regia will provide important protocols for the long-term conservation of this vulnerable species.Finally, this study shows the importance of multiple sites and long-term monitoring for successful restoration of rare species.

Plant Propagation
Silene regia seeds were collected in the Fall 2009 from a restoration site in Lawrence County, IL, U.S.A., containing plants that originated from an imperiled naturally occurring population in that county.Seeds were stored in a refrigerator at 4 C until stratification.To stratify, seeds were placed in filter paper (90 mm diameter, Whatman #1, Fisherbrand, Pittsburg, PA, U.S.A.) and wrapped in cheesecloth, creating a seed bundle.Seed bundles were soaked in Bonide Captan fungicide (50% wettable powder, 4.93 g/L) for 90 seconds.After fungicide treatment, seed bundles were placed in a plastic container (Rubbermaid 25 Â 15 Â 35 cm) with sand (Quickrete Companies, Inc., Atlanta, GA, U.S.A.) and sphagnum mixture (70:30% by mass).The sand/sphagnum mixture was moistened with 200 mL of distilled water.The plastic container was placed in a Fisherbrand Isotemp refrigerator (Pittsburg, PA, U.S.A.) set at 4 C. Seeds remained in the refrigerator for 8-10 weeks for cold, moist stratification following protocol in Flocca et al. (2004).
After stratification, a single seed was planted in each conetainer, 4 cm Â 21 cm with removable sleeves (Cone-tainer; Stuewe & Sons, Inc., Tangent, OR, U.S.A.).Cone-tainers were filled with soilless, high porosity potting mix and placed in a greenhouse for 9-14 weeks to allow for seedling growth.Four sets of seeds for planting in Spring 2010, Fall 2010, Spring 2011, and Fall 2011 were stratified, and seedlings were grown according to the above protocols.
Seedlings for the Spring 2010 and Fall 2010 plantings were grown with three treatments.One third of the seedlings received no fertilizer during growth in the greenhouse (control); one third received Peter's fertilizer (20-20-20) at 1.25 g fertilizer/L of water every other week (biweekly); one third received Peter's fertilizer (20-20-20) at 1.25 g fertilizer/L of water every week (weekly).For seedlings grown for the Spring 2011 and Fall 2011 plantings, an additional fertilizer treatment (slow-release) was used.Osmocote 4-month slow-release solid fertilizer 14-14-14 (Scotts-Sierra Horticultural Products Company, Marysville, OH, U.S.A.) was mixed into soilless media at 14.6 g fertilizer/L media at the time of planting the seeds into cone-tainers (slow-release).Before planting in the field, the seedlings were hardened off for 1.5-2 weeks outdoors in a greenhouse courtyard.Seedlings were exposed first to shade for 5-7 days and then to the sun for 5-7 days before planting.Initially, seedlings of S. regia form a rosette of leaves, and the number of green leaves in each rosette was counted before planting at sites.Stems had begun to elongate on some seedlings, and these were counted (see Table S1).Seedlings from the control and various fertilizer treatments were sorted randomly for planting in designated blocks at each site.

Study Sites
This research project was conducted at Prairie Ridge State Natural Area (PRSNA) in Jasper County, IL. Three prairie restoration sites (sites G, R, and W) within 2.5 km of each other were selected for the plantings of S. regia seedlings.Sites were characterized with soil nutrient analyses, soil moisture, and total cover.Differences between sites were found in most of these parameters.See Supplement S1 and Table S2 & S3 for additional information on soil sampling protocol and results of soil analyses and vegetation cover.The annual average temperature for Newton, IL was 12.6 C for 2010 and 12.9 C for 2011.Average monthly temperatures from April to October (i.e.growing season) were slightly warmer at 20.9 C in 2010 compared to 19.8 C in 2011.Annual precipitation was lower in 2010, with 108.9 cm, compared to 2011, with 141.5 cm.The growing season in 2010 was drier than in 2011 (72.6 vs. 92.1 cm).

Planting Protocol and Monitoring
At each site, a planting location was selected for the 2010 Spring and Fall plantings and another location for the 2011 Spring and Fall plantings.For fall plantings, vegetation at sites was cut with a rotary mower to facilitate access for transplanting due to heavy vegetation cover.Each planting location had two plots (one for spring and one for fall).Each plot included three blocks with 1 m 2 replication for each fertilizer treatment arranged in a randomized complete block design with 1-m wide pathways between blocks (Fig. 2).Seedling plugs were removed from cone-tainers and holes dug to accommodate the size of the plug.After planting, the soil was firmly replaced around the plug.To make holes for spring plantings, trowels or small spades were  S4).On 25 July 2022, an informal survey was conducted when the total number of plants (including potential recruits) by site and stems per plant were counted.In addition, seeds were collected in the fall of 2011 to test seed germination.See Supplement S2 for germination protocols and Table S5 for percentage of seeds that germinated.

Analyses
We used a univariate two-way ANOVA to evaluate differences in number of leaves within a few days before planting.Planting season (two levels), fertilizer (2010: three levels, 2011: four levels), and the interaction between the two were used as fixed effects.Separate ANOVAs were run for the 2010 and 2011 planting.The number of leaves was log-transformed to meet the model assumption of normal distribution of residuals.Back-transformed data are presented in the results.We used a Tukey HSD post hoc test to evaluate differences among fixed effects.
For year-to-year survival, overall survival, and number of plants with elongated stems, we used generalized linear models to evaluate differences (Bates et al. 2015).All response variables were modeled using a binomial distribution.Planting season (two levels), fertilizer (2010: three levels, 2011: four levels), planting site (three levels), and survey year (categorical, ranging from 2010 or 2011 to 2017), along with the interaction between planting season and fertilizer were used as fixed effects.We used Akaike's information criterion for small sample sizes (AICc) to compare models with differing fixed effects.A pairwise comparison of estimated marginal means with a Tukey's adjustment was used to compare differences among fixed effects.For these analyses, estimated marginal means from the emmeans package are presented in the results (Lenth 2022).All data analyses were run using R version 4.2.3 (R Core Team 2023).

Number of Leaves
Fertilizer treatment (F 2 = 70.53,p < 0.001), planting season (F 1 = 250.88,p < 0.001), and their interaction (F 2 = 3.21, p = 0.041) had a significant effect on the number of leaves for the 2010 planting.The plants fertilized with the biweekly and weekly fertilizer had approximately five more leaves than plants with no fertilizer (Fig. 3A).The plants used in the spring planting had more leaves than the plants used in the fall planting (Fig. 3A).
For the plants used in 2011, fertilizer treatment (F 3 = 345.96,p < 0.001), planting season (F 1 = 101.58,p < 0.001), and their interaction (F 3 = 29.14, p < 0.001) had a significant effect on numbers of leaves.Again, the plants that received biweekly and weekly had more leaves than the plants that received no fertilizer; however, the plants that had slow-release fertilizer had the most leaves (Fig. 3B).Liquid fertilizer resulted in plants having more leaves in the fall (approximately 18-21 leaves) compared to the spring (approximately 11 leaves) planting.The plants that received the slow-release fertilizer had over 10 more leaves than the liquid fertilizer treatments.

Year-to-Year Survival
For the 2010 planting, year-to-year survival ranged from 48 to 93%, with a mean of 71%.Planting season, site, and survey year influenced year-to-year survival and were all included in the top-performing model (Table S6).Fertilizer had almost no influence on year-to-year survival and had a cumulative weight of 0.247 (Fig. 4A).Cumulative weight is the sum of AICc weights (i.e. the predictive power as a proportion of all models), with lower weights representing lower predictive power.Seedlings planted in the spring were 1.25 times more likely to survive year-to-year than those planted in the fall (Fig. 4B).Site G had the highest year-to-year survival (Fig. 4C).Survey year after planting had a strong influence on year-to-year survival and was highest for the first year at 0.92, dropped to 0.78-0.81from 2011 to 2014, then dropped below 0.59 from 2015 to 2017 (Fig. 4D).
The 2011 planting had a mean year-to-year survival of 78%, with a range of 64 to 99%.Year-to-year survival was influenced by all variables, including the interaction between planting season and fertilizer (Table S6).Year-to-year survival probability was similar for all fertilizer treatments in the spring planting, while the biweekly fertilizer treatment had higher year-to-year survival compared to seedlings that received no fertilizer in the fall planting (Fig. 4E).Unlike 2010, year-to-year survival was highest for site R and lowest for site W (Fig. 4F).Similar to 2010, year-to-year survival probability was highest the first year at 0.99, then dropped to 0.85-0.89from 2012 to 2013, then dropped below 0.73 from 2014 to 2017 (Fig. 4G).

Overall Survival
Overall survival was influenced by all the tested variables in 2010 (Table S7).Overall survival was higher in the spring planting compared to the fall planting over all fertilizer treatments (Fig. 5A).Fertilizer treatment had little influence on overall survival for the spring planting; however, plants that received weekly fertilizer had higher overall survival compared to plants that received no or biweekly fertilizer for fall planting (Fig. 5A).Overall survival was highest for site G and lowest for site R (Fig. 5B).Overall survival probability steadily declined over the 8 years from 0.93 to 0.07 (Fig. 5c).
Similar to 2010, overall survival was influenced by all the tested variables in 2011 (Table S7).Overall survival probability was highest for the weekly fertilizer and slow-release fertilizer for the spring planting (Fig. 5D).All fertilizer treatments had higher overall survival probability compared to no fertilizer for the fall planting (Fig. 5D).Overall survival was highest for site R and lowest for site W (Fig. 5E).Again, overall survival probability dropped each year from 0.99 in 2011 to 0.18 in 2017 (Fig. 5F).
When the sites were revisited in 2022 to check the population status, site G and site R had 29 and 22 plants, respectively (from an original population size of 414 plants per site), while site W had no plants remaining.These surveyed plants had 1-6 stems each.

Plants with Elongated Stems
For the 2010 planting year, planting season, site, and survey year influenced the probability a plant had elongated stems the most (Table S8).Fertilizer had little influence on stem elongation and cumulative weight of 0.79 (Fig. 6A).Seedlings planted in the spring were 1.42 times more likely to have elongated stems compared to seedlings planted in the fall (Fig. 6B).Site G and site R had higher proportions of plants with elongated stems compared to site W (Fig. 6C).Very few stems were elongated the year of planting, but stem elongation probability was above 0.84 every year after except for 2016 (Fig. 6D).
For the 2011 planting year, the interaction between planting season and fertilizer, along with the survey year, had the strongest influence on stem elongation probability (Table S8).The planting site had no influence on stem elongation and had a cumulative weight of 0.60.The probability of stem elongation was highest for slow-release fertilizer, followed by biweekly and weekly fertilizer, and lastly, no fertilizer for the spring planting (Fig. 6E).For the fall planting, the biweekly fertilizer had a higher stem elongation probability than no fertilizer (Fig. 6E).Stem elongation probability did not differ among the planting sites (Fig. 6F).Stem elongation probability was below 0.10 for the first year and was above 0.86 for all following years (Fig. 6G).

Discussion
This study provides evidence for how fertilizer protocols to produce transplants and season of planting impact the survival and reproductive potential of Silene regia plants introduced for restoration.Fertilizer type had a strong and consistent effect on the number of leaves on seedlings.Plants grown with slowrelease fertilizer had the most leaves, followed by both liquid fertilizers and, lastly, those with no fertilizer.There was no advantage to using liquid fertilizer every week versus every other week.Our results showing increased seedling size with slow-release fertilizer agree with those for Penstemon tubaeflorus (Annis et al. 2014) and for Schizachyrium sanguineum (de Oliveira et al. 2022), although liquid fertilizer also increased seedling size for S. regia, but did not for P. tubaeflorus (Annis et al. 2014).While the number of leaves on plants before planting differed between planting seasons it was not a consistent effect.For the 2010 planting, plants used in the spring planting had more leaves for all fertilizer treatments than those in the fall planting.For the 2011 planting, the number of leaves was higher for plants used in the fall planting than in the spring planting for the liquid fertilizers.
While fertilizer had a strong and consistent effect on the number of leaves on transplants, it had an inconsistent effect on yearto-year survival, overall survival, and stem elongation.For the 2010 planting, fertilizer only influenced overall survival for the fall planting, with weekly liquid fertilizer being the highest.Fertilizer treatments had a greater effect on the 2011 planting.Fertilizer additions had higher survival and stem elongation rates compared to no fertilizer, although no particular fertilizer treatment always performed the best.However, plants treated with the slow-release fertilizer were more likely to have elongated stems before planting than any of the other fertilizer treatments (slow-release: up to 36%, all others: <4%).For the production of plants to use in restoration, slow-release fertilizer produced larger transplants at planting, but that increased size did not always result in greater survival or reproductive potential.It is important to note that slow-release fertilizer can remain in the soil after transplanting.While the impacts of these nutrient additions were not tested in this study, we believe the impacts would be low as many of the nutrients would have been absorbed or leached from the soil before planting.The impacts of these nutrient additions may vary based on the nutrient availability in various habitats.For S. regia restoration, fertilizer of some type should be used.
Planting season had a more consistent effect on year-to-year survival, overall survival, and stem elongation than fertilizer.The spring planting had higher year-to-year survival, overall survival, and plants with elongated stems for the 2010 planting.The 2011 planting similarly had higher survival and stem elongation rates in the spring compared to the fall, but only for a few fertilizer treatments.Page and Bork (2005) also found higher survival in spring than in fall for Achnatherum richardsonii (Richardson's needlegrass), although the reverse was found for Pseudoroegneria spicata (bluebunch wheatgrass).For three Carex species, Steed and DeWald (2003) reported greater survival with summer than with fall plantings.One potential reason spring may have been better for S. regia transplants is that soil moisture and precipitation were higher in the spring, which could facilitate better root growth and overall plant health.In addition, plants transplanted in the spring may have had more time to establish before the harsh winter conditions.For S. regia restoration efforts, spring would be the preferred season to plant over fall.
The importance of site suitability is demonstrated by the differences observed.When differences occurred, plants at site W often demonstrated the least success based on year-to-year survival, overall survival, and elongated stems.While site W often had the lowest available soil nutrients compared to the other sites, plant covers were similar between all sites, suggesting nutrients were sufficient for plant growth.However, site W also had a higher percent soil moisture in the Spring than other sites.This soil moisture supported crayfish activity, which disturbed soils and resulted in uprooted or buried transplants, which could have negatively impacted their survival.In addition, site management has the potential to impact reintroduction success, but management activities were applied similarly across all sites.All of these site differences stress the importance of implementing reintroductions at multiple sites.
Long-term monitoring is necessary to assess the success of reintroductions.Year-to-year mortality was initially low (<20%), which was similar to year-to-year mortality rates of 5-17% observed by Menges and Dolan (1998).However, mortality rates increased after 5 and 3 years for the 2010 (>40%) and 2011 (>30%) planting, respectively.In addition, overall survival was less than 20% by 2017 after 8 years of surveying.By 2022, only two of the three sites had plants remaining.In contrast, plants were quick to have elongated stems and presumably flowers, with stems elongating the first year after planting.Over 76% of the plants had elongated stems for all 6-7 years after planting.These findings are similar to the work by Menges and Dolan (1998), who observed that 75% of adult S. regia plants flowered.There was a dip in the proportion of plants with elongated stems in 2016 for the 2010 planting.This dip was likely due to the survey being done in May before stems may have begun to elongate.Had surveys ceased after 4 years, the usual length of monitoring after reintroductions (Godefroid et al. 2011), a more optimistic outcome would have been predicted.
Finally, natural recruitment from seed is crucial for the longterm survival of both common and rare native species (Albrecht et al. 2019).In the case of our study, recruitment is necessary to maintain the viability of these transplanted populations.The seeds collected in 2011 were viable and germination differed by site but ranged from 43% to 70%.In addition, 41 new plants were observed, with the earliest recruit being observed 3 years after planting and the majority being observed 4+ years after planting.Most of the new plants (approximately 80%) were observed near the spring planting plots.The seed viability and recruitment that we observed suggest the potential persistence of these S. regia populations.
Based on the findings of this study, planting in the spring and using fertilizer gave the best chance of creating viable populations of S. regia.In addition to planting season and fertilizer, planting in multiple sites increases the likelihood of population establishment.Finally, planting over multiple years can help mitigate the impacts of stochastic events (e.g.droughts and herbivory).Long-term surveys are necessary to determine the fate of transplanted populations, as populations may appear viable for the first couple of years before declining (Doyle et al. 2023).Incorporating these protocols should help the long-term success of S. regia restoration attempts and other rare species alike.

Figure 1 .
Figure 1.Photo of Silene regia transplant flowering at Prairie Ridge State Natural Area, IL.
used, but battery-powered drills were used to make holes for fall plantings due to the hard, dry soil conditions.Planting occurred on 19 May 2010 (12 plants/rep; total = 324); 3 October 2010 (10 plants/rep; total = 270); 6 May 2011 (12 plants/rep; total = 432), and 27 September 2011 (6 plants/rep; total = 216).Because of extremely dry conditions in the fall of 2010, 13-15 L of water was added to each replication on 3 October immediately after planting, and another 17 L was added on 13 October.No other seedling care was given after planting.To facilitate plant relocations for future monitoring, a small wire flag was placed next to each seedling after planting.Each area was fenced to exclude deer.After planting, counts were taken for the following parameters: number of live plants and number of plants with elongated stems.Stems elongate in preparation for flowering.Our observed rates of stem elongation were similar to observed rates of flowering byMenges and Dolan (1998).Counts were taken at monthly intervals in 2010 (June to October), 2011 (April to October), and 2012 (April to October except July and August because of drought conditions); twice a year in 2013 (May and July), 2014 (June and July), 2015 (May and July); and once a year 2016 (May) and 2017 (August) (Table

Figure 2 .
Figure 2. Diagram of the randomized complete block design for 2011.Each block contained a replication of all fertilizer treatments.Each replication was 1 m Â 1 m with 1 m walkways between replications and blocks.Treatments within each block were chosen at random.2010 had the same design without the slow-release treatment.

Figure 3 .
Figure 3. Mean number of leaves by season and fertilizer treatment of plants used in the (A) 2010 and (B) 2011 planting.Back-transformed data are presented.A Tukey HSD post hoc test was used to evaluate differences among fixed effects.Ages of plants when the leaves were counted: Spring 2010-80 days; Fall 2010-110 days; Spring 2011-67 days; Fall 2011-96 days.Error bars indicate AE SE.

Figure 5 .
Figure 5. Overall survival probability as a function of fertilizer, season, site, and survey year for the 2010 (A-C) and 2011 (D-F) planting years.Estimated marginal means are presented from top models in TableS7.Letters represent differences among fixed effects determined using Tukey's adjustment.Error bars indicate AE SE.