Abstract
Rare plants garner significant conservation attention, but many have knowledge gaps associated with their life histories. This missing information presents a substantial hurdle for applied conservation. Houghton’s goldenrod (Solidago houghtonii Torrey & A. Gray [Asteraceae]) is a Great Lakes endemic perennial that is often locally abundant but is limited to a narrow region in Michigan, New York, and Ontario. This species is federally listed in the US as threatened, state listed as threatened in Michigan and endangered in New York, and is a species of special concern in Canada. There may be enough viable S. houghtonii populations to meet the US federal recovery criterion, but more data are needed regarding long-term persistence within and across populations, especially related to successful sexual reproduction—key in most seed plants’ long-term persistence. As part of a range-wide study on the species, we performed greenhouse and field experiments to evaluate the effects of substrate moisture and texture on germination success. In both experimental contexts, we found that S. houghtonii germination significantly increased with increased substrate moisture and smoother substrate texture. In addition, a post hoc exploratory analysis on the effects of disturbance on germination showed higher germination in plots with higher disturbance levels, though this trend was not statistically significant. These results add to the limited life history information available for S. houghtonii. Our findings also suggest that substrate texture may be an easily characterized habitat variable that predicts germination for other species because of its relationship to consistent moisture contact for seeds.
Weber JB, Leopold DL. 2023. Substrate moisture and texture affect germination in Houghton’s goldenrod (Solidago houghtonii), a federally protected Great Lakes endemic plant. Native Plants Journal 24(1):33–43.
Solidago houghtonii (Houghton’s goldenrod [Asteraceae]) inflorescence.
Conservation of rare species relies on understanding niche requirements, life history bottlenecks, and specific threats to persistence; however, major knowledge gaps exist for these critical factors in many rare plant species (Havens and others 2014; Aronne 2017). These gaps present a substantial obstacle in conservation planning (Knapp and others 2020; Nic Lughadha and others 2020). Furthermore, much research on plant species’ traits and niche requirements focuses on mature plants (Jiménez-Alfaro and others 2016), even though rare plants may be more strongly limited by processes related to sexual reproduction, such as seed germination and seedling establishment (Burmeier and Jensen 2008). And many species experience shifts in niche requirements throughout their life histories: Seeds and seedlings often have needs that differ from mature plants, for example (Maguire 1973; Pironon and others 2018). Understanding regeneration requirements, especially the conditions necessary for sexual reproduction, is crucial for informed, long-term conservation (Grubb 1977; Donohue and others 2010).
Houghton’s goldenrod (Solidago houghtonii Torrey & A. Gray [Asteraceae]) is a hexaploid, perennial Great Lakes endemic that is often locally abundant within a narrow region along the Niagara Escarpment (USFWS 2011). Approximately 90 occurrences (loosely, populations) are in Michigan and Ontario with one occurrence in western New York (USFWS 2020) (Figure 1). Solidago houghtonii is federally listed as threatened in the US, state listed as threatened in Michigan and endangered in New York (USFWS 2020), and is a species of special concern in Canada (COSEWIC 2005). The US federal recovery criterion states that at least 30 of the most viable populations must occur in protected habitat (USFWS 1997). While the species is generally uncommon, there may now be enough protected S. houghtonii populations to meet the federal recovery criterion, and it may be appropriate to consider the species for delisting from US federal protection. However, more data are needed regarding long-term viability or stability within and across populations, and there continues to be scant knowledge beyond anecdotal information on the species’ regeneration niche.
Range extent for Solidago houghtonii (top) and Michigan field site locations for germination field experiments (bottom). An additional field site was established in the single New York population seen in the range map here. Base map tiles are from ©Stamen Design, used under CC BY 3.0.
Solidago houghtonii readily produces asexual clones via rhizomes but requires insect cross-pollination for sexual reproduction; successful sexual reproduction in S. houghtonii is likely relatively infrequent due to the multiple requirements from pollination to seedling recruitment, all in dynamic habitats (USFWS 1997). While assessing population viability is fundamental in managing rare species and a component in the S. houghtonii Federal Recovery Plan (USFWS 1997; Jolls and others 2015), S. houghtonii’s abundant clonal offspring and unknown dormancy habits make accurate estimations of population growth rates difficult (Menges 2000; Morris and others 2002). In lieu of this, measuring rates of important demographic transitions—such as germination—can provide a proxy for reproductive success and estimates of population stability (Jiménez-Alfaro and others 2016; Aronne 2017). More specifically, successful germination represents a culmination of processes involving genetic recombination and seed dispersal—potentially increasing genetic diversity and new site colonization, respectively—both key in long-term persistence of many seed plants.
Dynamic and abiotically stressful habitats present many challenges for successful germination and establishment (Keddy and Constabel 1986). Solidago houghtonii primarily occurs in transitional areas between aquatic and terrestrial environments, that is, wetland and dune-swale systems. Its populations are frequently exposed to natural and human-driven lake flooding, shifting substrate, and other abiotic stresses (USFWS 1997, 2020). Substrate moisture and flooding regime are often the most important factors regulating germination in wetland plant species (Moore and Keddy 1988). Specifically, relatively constant moisture is often an important germination cue, and the depth, duration, and frequency of flood events may restrict the pool of species able to successfully germinate (van der Walk 1981). Seed contact with moisture can also be influenced by substrate texture, so substrate type can indirectly influence germination. Certain substrate textures may also protect seeds from disturbances such as flooding or sudden desiccation (Keddy and Constabel 1986; Pinno and others 2017).
We compared S. houghtonii germination across typical moisture regimes and substrates. We predicted that more consistent moisture treatments would result in the highest germination success, and that germination would be highest in rough substrates containing organic matter (for example, moss or litter) because of their ability to protect seeds from desiccation and disturbance. We tested these predictions using complementary greenhouse and field experiments. Because disturbance events such as flooding, shifting substrate, or trampling by human activity likely have an impact on germination, we performed a post hoc exploratory analysis on the effects of disturbance on germination success. These results add to the limited life history information available for S. houghtonii and, combined with additional habitat data, will help to inform conservation prioritization across the species’ range.
MATERIALS AND METHODS
Seed Collection
In October 2015, we collected seeds from 27 S. houghtonii populations across the extent of its US range (26 in Michigan, 1 in New York; federal permit No. TE71508B-0). At each site, we collected 1 infructescence, each, from 5 to 10 widely spaced individuals, depending on the extent of the occurrence. Mature infructescences were kept in cold, dry storage until sorting. Some seeds exhibited insect damage at the time of collection, but only firm, intact seeds were used for experimentation. We randomly sorted each population’s seeds into sample groups of 25 seeds, sampling from all collected infructescences, and cold-moist stratified them in 60 ml (2 oz) of sterile peat—lightly moistened with tap water—at 4 °C (39.2 °F) for approximately 120 d. From each population, 10 sample groups were randomly selected for greenhouse (ex situ) experiments at State University of New York College of Environmental Science and Forestry (SUNY ESF) in Syracuse, New York; the 10 remaining sample groups were used in field (in situ) experiments at the original collection sites.
Greenhouse Experiments
We tested the effects of substrate and moisture on germination with 2 controlled greenhouse experiments during May 2016. Greenhouse temperatures were set to average early growing season temperatures in coastal northern Michigan, the core of S. houghtonii’s range: 18 °C (64.5 °F) during the day and 9 °C (48.2 °F) at night (May and June averages; NOAA NCEI 2016). No grow lights were used to extend ambient outdoor light conditions. For each experiment, we filled 10 cm × 10 cm (4 in × 4 in) plastic pots with substrate and spread 1 sample group of stratified seeds (25 seeds from 1 population), with stratification medium, over the surface of each pot (Figure 2), for a total of 27 pots (that is, populations or experimental units) per treatment. We randomly placed pots in trays by treatment and shuffled pots within trays daily.
Experimental setup. Greenhouse moisture experiments from the same source population of seeds. From left to right: high/frequent, low/frequent, high/infrequent, and low/infrequent treatments (A). Cotyledon emergence used for germinant counts (B). Greenhouse substrate experiments from the same source population of seeds. Clockwise from top left: gravel, marl, sand, moss, litter, soil. Note the visible differences in water behavior and seed contact across substrates, despite the presence of stratification medium (C). Field experiments: seeds and stratification medium spread in 10 cm × 10 cm (4 in × 4 in) square directly south adjacent to plot marker (D).
A typical dune-swale system along Lake Michigan. Note the visible PVC tube plot marker from this study and the narrow band of open habitat bounded by the lake on one side and dense forest on the other.
Mature flowering individual of Solidago houghtonii growing on sand in a dune system along Lake Huron.
Typical intact Solidago houghtonii seed with pappus. One cm on ruler for reference.
For the moisture experiment, we used a factorial design with 2 levels of 2 factors: volume and frequency of watering. This design resulted in 4 treatment combinations intended to simulate precipitation ranges for northern Michigan in May and June (NOAA NCEI 2016): 25 ml (0.85 oz) and 50 ml (1.70 oz) of water for low and high treatments, respectively, and watering every other day or every day for infrequent and frequent treatments, respectively. Pots were filled with commercial propagation mix (Sungro Sunshine Mix #5 containing lime, perlite, peat moss, and silicon; Agawam, Massachusetts). We counted germinants in each pot every day before watering. Seed germination was defined as cotyledon appearance, given that radicle emergence was difficult to reliably find. We counted germinants until no new germinants appeared for 2 d, approximately 3 wk total.
We used 6 substrates in the substrate experiment: marl, sand, soil, moss, gravel, and litter. These materials represent typical substrates on which S. houghtonii can be found in situ, and actual treatment substrates used in the greenhouse experiments were matched as closely to in situ conditions as possible. We collected marl (calcium carbonate deposits) and sand from sites adjacent to S. houghtonii populations in New York and Michigan, respectively; we used commercial propagation mix (Sungro Sunshine Mix #5) as a soil substrate to mimic the organic soils of S. houghtonii habitats; we gathered sheets of moss in Central New York of similar life-forms and calciphilic genera to those found in S. houghtonii habitats; we purchased limestone, medium gravel (EN ISO 14688-1 gravel size) to replicate the gravel flats found in some areas around northern Lakes Michigan and Huron, which have similarly sized particles originating from primarily limestone bedrock; and we purchased dried straw, which has a similar physical structure to litter wrack common along the Great Lakes’ coastlines. We watered pots with tap water until saturation every other day and counted germinants daily as described above.
Field Experiment
The field germination experiment took place in June 2016 at 16 population sites across S. houghtonii’s range, 15 in Michigan and 1 in New York (Figure 1). Solidago houghtonii is primarily found in open sites with calcium-rich substrates: Great Lakes coastal dune-swale systems, coastal alvars, gravel flats, and marl fens (USFWS 1997). While the New York field site was an inland marl fen, our Michigan field sites were dominated by dune-swale systems and gravel flats—both narrow bands of habitat between Lakes Michigan or Huron and dense conifer forest.
At each site, we examined up to 30 plots, 0.25 m2 (2.7 ft2) in size, using stratified random sampling (loosely, by distance from shoreline) for a broader study of S. houghtonii niche characteristics. We randomly selected up to 10 of those study plots for germination experiments (n = 148). We spread 1 sample group of stratified seeds (with stratification medium) in a 10 cm × 10 cm (4 × 4 in) area directly south but adjacent to the plot’s center (Figure 2). We visited sites weekly, for 4 wk, to record number of germinants, associated substrate of each new germinant, and surface moisture levels (as volumetric water content, m3/m3, using a Decagon Devices GS3 sensor; Pullman, Washington). Reliable counts of non-germinated seeds were nearly impossible without disturbing plots or seed substrates.
Basal rosette of Solidago houghtonii growing on rocky shorelines along Lake Huron.
Statistical Analysis
We compared differences in germination among greenhouse moisture and substrate treatment groups using binomial generalized linear models in R (v4.1.3; R Core Team 2022) and the proportion of seeds germinated as the response variable (for each experiment, n = 27). Moisture experiment analysis also included an interaction term between volume and frequency of watering. We conducted pairwise comparisons for treatments in both experiments using Tukey’s HSD (α = 0.05; R package “multcomp” 2008).
Since a high number of plots had no germination in situ, we used zero-inflated binomial generalized linear mixed models (R package “glmmTMB” 2017) to analyze the relationship of field germination after 4 wk as the response variable and mean plot moisture (averaging moisture readings from the season) as a continuous predictor variable (n = 147) (Table 1). We included population as a random effect. Given the results from the greenhouse substrate experiment and uneven field substrate group sample sizes, we also used zero-inflated binomial mixed modeling to compare differences in germination between “rough” (gravel, litter, and moss) and “smooth” (marl, sand, and soil) substrates, with the proportion of seeds germinated in situ after 4 wk as the response variable and population as a random effect (see Table 2 for group sample sizes). We conducted pairwise comparisons for substrate groups using Tukey’s HSD (α = 0.05; R package “multcomp”). Because field substrate and moisture are generally associated, we also created combined field substrate and moisture models, including a model with an interaction term, and used Akaike’s Information Criteria (AIC) to compare which models best predicted germination in the field (Table 3).
Descriptive statistics for germination success (proportion) by moisture treatment in greenhouse experiments and continuous moisture variable in field experiments.
Descriptive statistics for germination success (proportion) by substrate treatment in greenhouse and field experiments.
Model comparison table for S. houghtonii germination zero-inflated generalized linear mixed models using Akaike’s Information Criteria (AIC).
Post Hoc Disturbance Analysis
To investigate potential effects of disturbance on germination as a post hoc exploratory analysis, we developed a disturbance index to rate each field plot based on 4 criteria (aspect, substrate, acute flooding potential, and public access) and 3 ranks for each criterion (low, medium, high) (Table 4). For each field plot, a value of 1, 2, or 3 (for low, medium, or high, respectively) was assigned for each criterion, and values were averaged to determine the disturbance index value for that plot. Because final index values were not truly continuous data, we created 3 bins of equal intervals for analysis: plots with disturbance index values from 1.00 to 1.67 were rated as having “low” disturbance overall, plots with values from 1.68 to 2.33 as “medium” disturbance, and plots with values from 2.34 to 3.00 as “high” disturbance. We compared differences in germination between disturbance-level groups using the same methods as the field treatment groups above: zero-inflated binomial mixed models (R package “glmmTMB”), with population included as a random effect and pairwise comparisons via Tukey’s HSD (α = 0.05).
Disturbance ranking criteria and levels.
RESULTS
Greenhouse Experiments
In the greenhouse moisture experiments, the effects of water volume and frequency were not independent (P = 0.01): The interaction between water volume and water frequency was significant. Significant main effects were found in low (frequent compared with infrequent, P < 0.001) and infrequent (low compared to high, P < 0.001) levels. In short, the lowest-moisture treatment resulted in the lowest germination (Table 1). Notable variability occurred across populations (replicates) within treatments (Figure 3).
Mean (± SE) proportion of seeds germinated. Proportion of seeds germinated in greenhouse (top) and field (bottom), substrate (left) and moisture (right) experiments. Dots each represent 1 experimental unit: populations in greenhouse experiments and plots in field experiments. Letters above bars indicate statistically significant groupings via Tukey’s HSD, at α = 0.05; groups that share a letter are not statistically different from one another. The moisture field panel shows predicted germination by substrate texture (smooth compared to rough) across the moisture gradient.
In the greenhouse substrate experiments, smoother substrates (marl, sand, and soil) had significantly higher germination than did rougher substrates (moss, gravel, and litter), with the highest (40%) and lowest (15%) mean proportion of germination on marl and litter, respectively (Table 2). All pairwise comparisons between smooth and rough substrates (for example, sand compared to moss, marl compared to litter) were significantly different. Again, notable variability occurred across populations within treatments (Figure 3).
Field Experiments
Field experiments had much lower overall germination success than the greenhouse experiments. Even so, mean substrate moisture as a continuous predictor variable was significantly positively associated with in situ germination success (P < 0.01): Higher moisture levels resulted in higher germination (Figure 3). Also, smoother substrates had significantly higher in situ germination (P < 0.01). The highest (6.2%) and the lowest (1.5%) mean proportion of germination occurred on sand and moss, respectively (Table 2).
When modeled together, the interaction of substrate and moisture was not significant (P = 0.697) but the difference in germination success between smooth and rough substrates was most pronounced at the highest moisture levels (Figure 3). In addition, of the models used for field germination analysis, the best model for predicting germination success included both substrate and moisture (Table 3). Model comparisons showed that including substrate and moisture data (either or both) better predicts germination success than the null model with only the random effect of population (Table 3).
Disturbance Analysis
Plots in the lowest disturbance group (n = 15) also had the lowest germination success (3.5%), while plots in the highest disturbance group (n = 100) had the highest germination (5.3%). Plots in the moderate disturbance group (n = 34) had 4.8% mean germination (Figure 4). Despite the apparently positive relationship between germination and disturbance in these data, differences between groups were not statistically significant (P > 0.2 for all comparisons).
Mean (± SE) proportion of seeds germinated per disturbance level in field experiments. Dots represent plots across 16 field sites.
DISCUSSION
Moisture regime and substrate texture were both clearly important for successful S. houghtonii germination in greenhouse and field experiments: Higher moisture and smoother substrates resulted in significantly higher germination. These results only partially support our original predictions of more consistent moisture and rougher substrates being most associated with successful S. houghtonii germination.
Moisture Effects
Like virtually every other seed plant, S. houghtonii seeds depend on a minimum threshold of water imbibition as a germination cue (Baskin and Baskin 2014), and since moisture is an especially prominent factor in wetland species’ germination (Keddy and Reznicek 1986), this variable is plainly important in S. houghtonii germination success. In both our greenhouse and field experiments, germination was significantly lower in the driest conditions. In situ hydrology, especially in S. houghtonii’s driest habitats (for example, Great Lakes dunes) likely limits germination success.
Increased variability in moisture availability across S. houghtonii’s typical substrates (sand and moss) and habitats (dune and fen)—from shifts in precipitation amounts or timing because of climate change—will likely affect S. houghtonii germination success. Decreased access to moisture cues during germination windows, or too much moisture (such as acute flooding) before establishment, will lower germination success (Walck and others 1997; Fay and Schulz 2009; Walck and others 2011). However, meaningfully characterizing moisture at the level of a seed is extremely challenging—especially in the field—so more precise statements are nearly impossible.
Substrate Effects
The differences in S. houghtonii germination between substrates, especially in the greenhouse experiments, appear to be related to substrate texture, which in turn relates to moisture contact for germination: Smooth textures provide more uniform contact with a saturated surface (Harper and Benton 1966). While rougher substrates, such as moss, litter, or gravel, may provide protection from displacement for seeds in situ, this protection may be negated without consistent moisture contact. Previous research appears to confirm this relationship of substrate, moisture contact, and germination niche, although no studies seem to include substrates beyond soil and discuss only soil particle size (Harper and Benton 1966; Keddy and Constabel 1986; Pinno and others 2017). In addition, substrate texture requirements may change over a plant’s life span: Smooth textures, such as marl or sand, may provide adequate moisture contact for germination but afford limited nutrient availability for establishment, growth, and survival (Eckhart and others 2017; Pinno and others 2017); or, as mentioned in our original hypotheses, smooth substrates may not provide in situ protection from flooding or rapid desiccation. Differences between this study’s greenhouse and field germination—generally somewhat successful and quite low, respectively, across substrates—suggest that while substrate texture is important, it may not be the main factor in S. houghtonii’s field germination and establishment success. However, our results show that substrate texture is an important predictor of germination—likely attributable to its relationship to moisture contact for seeds—and because substrate is much easier to characterize in the field than is moisture, it may be an even better variable than moisture on which to focus for seed-related applied conservation.
Disturbance and Germination
Solidago houghtonii often occurs in highly dynamic dune-swale habitats along northern Lakes Michigan and Huron (USFWS 1997). Disturbance is common in these habitats and can include shifting substrate (sand), acute flooding, intense fall storms and winter ice scour, and human activities such as beach recreation and off-road vehicle use (USFWS 2011; Leopold and Weber 2019). In fact, most protected populations of S. houghtonii occur in Michigan state parks or state forests with few restrictions on recreational use of coastlines (Leopold and Weber 2019). The dynamic nature of Great Lakes coastal habitats likely has major impacts on year-to-year survival of newly established plants (Keddy and Reznicek 1986). As such, disturbance events represent stochastic limits on S. houghtonii germination and establishment, thus limiting reliable conservation planning. While these disturbances represent challenges for informed management, disturbance has long been associated with success for certain plant communities and functional groups. Disturbance processes often eliminate more competitive woody plants, opening substrates for light-seeking new colonizers, as suggested by the Competitive Release Hypothesis (Grime 1977; Keddy 1990; Jutila and Grace 2002).
CONCLUSIONS
Substrate moisture and texture are influential factors in the germination success of S. houghtonii. Because substrate texture is easy to evaluate in the field, it provides an important tool for evaluation and prioritization of vulnerable S. houghtonii populations. Knowledge of a plant species’ regeneration niche—the resource space required for germination or clonal reproduction—is critical for long-term conservation, especially as a regeneration niche may help to differentiate imperiled species and populations in terms of relative success, abundance, and persistence across space and time (Grubb 1977; Aronne 2017). And isolating key, easy-to-measure variables involved in reproductive success will streamline conservation of rare or declining species.
The effects of climate change may be particularly important to monitor in protected plant species as we consider how regeneration niches shift in the future, since changes to temperature and precipitation will affect substrate moisture characteristics (Walck and others 2011; Albrecht and others 2020). For example, our cold-moist stratification period was 120 d, 30 d longer than the longest known cold-moist stratification for germination experiments with S. houghtonii, and our germination success was much higher (our maximum was 40% compared to the earlier maximum of 19% in Jivoff 2014). Climate warming will shorten natural stratification times and could strongly limit S. houghtonii germination (Bandera and others 2019; Finch and others 2019; Albrecht and others 2020). In addition, climate warming may disproportionately negatively affect species with small ranges in narrow climatic zones, such as the narrow band of lake-influenced coastal habitat occupied by S. houghtonii, exacerbating survival and regeneration challenges (Ohlemüller and others 2008).
Future seed ecology studies of S. houghtonii or other rare Great Lakes dune-swale species should test the effects of burial depth and timing on germination and establishment in sandy habitats; while previous S. houghtonii experiments suggest germination cannot happen with burial since it needs light to cue germination, no published work exists on this topic to date (USFWS 1997). More broadly, understanding plant survivorship and population persistence within disturbance contexts is critical to long-term conservation prioritization and planning. Coupling these germination trials with additional field data, such as vegetation community, flooding duration and frequency, will expound on factors most associated with S. houghtonii abundance and flowering, and such information will help to inform the species’ future conservation prioritization.
ACKNOWLEDGMENTS
This work was conducted with funding from the US Fish and Wildlife Service Great Lakes Restoration Initiative. Permission to collect seeds and establish field plots was given by the USFWS, Michigan DNR, The Nature Conservancy, and Bergen Swamp Preservation Society. Many thanks to John J Wiley Jr (USFWS) for project development advice, Terry Ettinger (SUNY ESF) for greenhouse access and guidance, Lindsay St Windsor and Luria Lee for their help collecting and processing data, and Alex Petzke and Sara Scanga for reading previous versions of this manuscript.
Footnotes
Photos by Justine B Weber
REFERENCES
Jump to section
Related Articles
Cited By...
- No citing articles found.