Barto’s Nice Experiment Seed Set

Hi Flog following!

I’m excited to have worked in the Echinacea Project’s lab at the Chicago Botanic Garden for the past few days. While I was here, I worked on the seed set portion of my summer 2017 REU project (which I have now named Barto’s Nice Experiment), and I caught up with many friends from the field season. It’s definitely been a fun week for me in the Chicago area!

At the end of the summer, the team collected my experimental Echinacea heads from the Nice Island remnant in Minnesota after I left. When I came to the garden with Tracie on Thursday of last week, I began dissecting them. I started by separating all rows from each other, but I quickly realized I was only able to accurately distinguish the odd rows (which had painted bracts). To work efficiently, I categorized all achenes into 4 groups based on where they came from in each head: Row 1, Row 3, Row 5, Row 7, or Even Row. Between Thursday, Friday, and the first half of today, I cleaned my 21 experimental heads. Each of the odd rows were put into their own baggie and attached to an x-raying sheet. With the guidance of Tracie, I was able to capture images that show the fullness of all of my odd-row achenes. With this data, I can create a GLM in R like I did with my pollination data from the summer and model which experimental variables (row within the capitulum, style age, and pulse/steady pollination treatment) affected the seed set in my experiment.

For now, I am going to count the full/partially full/empty achenes in my x-rays and get ready to return to Arkansas tomorrow.

Ashley working in the Echinacea Project lab at the Chicago Botanic Garden.

Setting up the achenes for x-raying! Each of the baggies here has one row’s worth of achenes.

Final x-ray product for one row of achenes.

Nina’s Externship Report

Hello, Nina here! These past few months, I was an extern with the Echinacea project, looking specifically at data collected last summer about Echinacea success and surrounding plant diversity. My findings are summarized in the report that I’ve attached here.

Trevor’s Mini Report: Echinacea Externship 2017

Trevor Hughes

Prof. Wagenius

8 December 2017

Field Site Burning Effects on Liatris and Solidago Species: Year 2

Background Info:

       Habitat fragmentation threatens the livelihood of many remnant prairie populations in the midwestern United States. This is the result of various biological mechanisms such as limited genetic diversity and transformation of the surrounding ecosystem, both of which yield direct effects on the plant species coexisting there (Newman and Pilson 1997, Saccheri et al. 1998). Through either modification potential mates location as well as other changes in pollinator behavior, the reproduction of these species is left especially vulnerable to change (Wagenius 2006). These reproduction patterns are measured in what is commonly referred to as seedset, or the percentage of fertile achenes (those that possess an embryo) out of a certain grouping of total achenes; therefore, seedset effectively displays the relative amount of achenes that are fertile on any given inflorescence.

       Previous research reveals that both pollinator activity and respective reproduction of native plant species decrease with habitat fragmentation; as well, many of these studies focused on select species, namely the Echinacea angustifolia, a common plant species often found in midwestern prairie remnant habitats (Wagenius 2006). Differing from prior research, this study focuses on how a contemporary trend—the burning of a specific remnant prairie habitat—could be affecting the reproductive activity of remnant prairie plant species. Expanding beyond Echinacea, this study focuses on two other key prairie species: Liatris and Solidago. By examining two different sides (east and west) of a specific remnant prairie site where the east and west sides are burned interchangeably on a repeating pattern: east side burned, west side burned, and no side is burned. This report analyzes the findings concerning the second year of the study, where no field site was burned. While the specific findings discussed will be limited to this year’s data, with next years data, the study hopes to aggregate all years to yield some conclusions as to the effects of field site burning on reproduction of prairie plant species. Aside from investigating the possible effect that field-site burning has on the seedset of the plant species in the area, it also hopes to be able to provide comparable findings between Liatris, Solidago, and Echinacea so that findings regarding habitat fragmentation can be generalized to more prairie plant species in the future.


       Liatris and Solidago were gathered at the Staffanson site within the Douglas county in western Minnesota (centered near 45°49’ N, 95°43’ W). They were collected using a randomized method, with 90 Liatris and 74 Solidago inflorescences collected in total. However, the Liatris and Solidago species were collected from both the East and West sides of the Staffanson site, but a blind-procedure was used to ensure this was not known during the data collection or analysis process. It should be noted that while the West field site was burned last year, no field site was burned this year in preparation for burning the East field site and continuing the cycle in 2018.

       Following the collection process, the Liatris and Solidago inflorescences were randomized using excel software and put in their appropriate order, divided by species. This prevents bias at multiple steps during the course of data collection, including cleaning, counting, and x-raying. Beginning with Liatris, a bag was selected according to the random order and all heads were removed from the inflorescence to be cleaned. Next, thirty achenes were selected randomly using a grid system under the container to be used as an x-ray sample for that inflorescence. The remaining achenes were separated from any chaff and placed in an envelope for future reference and potential research. This process was repeated with all 90 Liatris samples. Additionally, beginning with the 50th randomized Liatris sample and continuing until all remaining Liatris samples were cleaned, the number of achenes in the top, middle, and bottom head were also recorded in a spreadsheet. If there was no clear middle, the achene closest to the bottom relative to the midpoint was chosen.

       The Solidago process was quite similar. The bags were selected in a pre-randomized order and the inflorescence was removed to be cleaned. However, due to the considerable number of achenes on most Solidago inflorescence, it would be impractical to attempt to use all achenes for cleaning; therefore, I simply shook each inflorescence for 10 seconds and used those achenes for further analysis. From there, 30 achenes were randomly placed into a small plastic baggie for x-raying whilst other achenes were separated and placed into an envelope for further comparison. This process was repeated for all 74 solidago samples.

       After the cleaning was completed, the baggies were placed on sheets to prepare for x-raying. Twenty samples were x-rayed at a time and there scans were recorded. Lastly, the x-ray samples were counted for the number the number of achenes in each sample that actually contained a fertile embryo; in this way, the seed-set of the Liatris and Solidago achenes was recorded so that comparisons between the east and west Staffanson field sites could be made.


       All data was aggregated and analyzed with use of various statistical tests on the R platform; all figures and/or tables were also made on this same platform. Data regarding the top/middle/bottom breakdown of Liatris along with the seedset of Liatris and Solidago on the east and west sides of the Staffanson site were analyzed for potential findings.

Figure 1: Location of Head vs. Size of Head in Tops/Middles/Bottoms of Liatris

       The tops of the Liatris species were found to have the greatest number of achenes with a mean number of achenes of 24.367, middles followed with a mean of 21.667, and bottoms yielded the lowest mean at 20.867 (see Figure 1). A bootstrap test was conducted between the means of the tops and middles, generating a p-value of 0.019, meaning that there is only a 1.9% chance that the difference in size of head betweenst tops and middles would be found by chance simply by sampling randomly from inflorescences (see Figure 2). While a p-value was not calculated for the difference in size of head amongst tops and bottoms, it is speculated that this analysis would yield a even lower p-value than tops and middles due to the greater difference in mean. While a bootstrap test was run between middles and bottoms,  this yielded a high p-value and indicated there was a large possibility that the difference between middles and bottoms of Liaris could be found by chance from sampling randomly and was therefore not determined to be an meaningful finding.

Figure 2: Bootstrap Test for variance between Tops and Middles of Liatris

       The data collected on seedset for the 90 Liatris samples was aggregated and eventually divided between East and West lines of the Staffanson site. The mean seedset was calculated for each side and are compared below (see Figure 3). The East meen seedset of Liatris was 35.2%, which towers over the West’s 20.6%.

Figure 3: Side of Staffanson Site vs. Mean Seedset of Liatris

       A bootstrap test was run on this dichotomy to determine the confidence intervals of these findings as well as the p-value. The ablines were found to fall outside of the normal distribution, yielding a p-value of 0.1%, indicating that this difference has a very little possibility of resulting from chance and that the difference in seedset between the East and West is quite noteworthy and telling of other possible differences between these two sites. The confidence intervals were not found to overlap, strengthening the certainty of my findings (see Figure 4).

Figure 4: East and West Sides of Staffanson vs. Mean Seedset of Liatris with 95% C.I.

       Identical tests were conducted on the 74 Solidago samples and respective seedset data, but different findings were discovered. The seed-set data collected for Solidago was divided between the east and west sides of the Staffanson site. The mean seed-sets of each side were calculated, revealing to be 35.9% for the east and 32.7% for the west (see Figure 5). This demonstrates that the east had a higher seed-set, but not at the same large interval seen with the Liatris species. Confidence intervals and a p-value were also determined for each side. The p-value was revealed to be 36.45%, which indicates the potential for this difference in seed-set to result a large portion of the time from natural variation, and not possibly due to the difference in side of the Staffanson site (see Figure 6). Supporting this large p-value, the confidence intervals also overlap, indicating that the variables could potentially swap and vary greatly in repeating this study, generating further doubt on the implications of this finding.

Figure 5: Side of Staffanson Site vs. Mean Seedset of Solidago

Figure 6: East and West Sides of Staffanson vs. Mean Seedset of Solidago with 95% C.I.


Firstly, the data regarding the tops/middles/bottoms of Liatris demonstrate a definite correlation among location and size of head. Based on the low p-value, it can be determined that the tops have the greatest number of achenes, compared to the middles and bottoms, but whether or not there is a significant difference in the number of achenes betweenst middles and bottoms cannot be directly determined from this data set. While this data indicates a trend between location and size of head (with greatest numbers of achenes found among the top heads), more samples should be collected and  analyzed to strengthen this finding, specifically with regards to the difference in size between middles and bottoms.


Before the data from this years seedset among the east and west sides of the Staffanson site can be analyzed, the findings from last year’s study (when the West side was burned), must be discussed. The 2016 data indicates that the West had a higher mean seedset than the East for Liatris, but there was little to no correlation between the east and west with Solidago. This finding is interesting for a couple of reasons. For one, this directly contrasts with my data from this year (where no sides were burned), as now the east had nearly a 75% increase in seedset compared to the West. Interestingly, the east side now has the larger seedset, meaning the sides of the field swapped by a large margin. This also contrasts with what one would expect with a field burning, which generally is speculated to stimulate growth, achene development, and seedset. However, it appears from this data set that the burning possibly hinders achene development and seedset, but this cannot be completely determined from this data set. The data from 2018 (when the East side is burned), would be vital for drawing this conclusion.

Interestingly enough, while the data sets contradict themselves on these grounds, they agree on one important point: field site burnings have a greater impact on Liatris than Solidago. This raises questions as to what biological mechanism within each species accounts for this difference. Once again, data from 2018 would also be vital for making a definitive conclusion as to why field site burning seems to have a lessened impact on the Solidago species.

Lake Forest College Intern: Marisol

Marisol, our 2017 Lake Forest College intern!

This fall the Echinacea Project had a Lake Forest College intern, Marisol. In Marisol’s 16-hour internship over 4 weeks, she accomplished a lot! Her primary goal was to assess if the seed counter will be useful in our ACE protocol. Currently, volunteers scan achenes and then count them with our counting software online. The seed counter could potentially remove these steps and make the process more efficient.

Marisol did a few experiments with the seed counter. She determined the best sensitivity and speed settings for counting Echinacea achenes, by running packs of achenes through the seed counter multiple times and comparing those counts to her manual counts. She found that the most accurate sensitivity and speed is 8 and 70, respectively.

Marisol also determined the types of achenes that get counted and the types of achenes that often get missed. Usually, achenes that are really small, thin, and broken don’t get counted at all. Broken achenes that are still pretty large often get counted (about 2/3 of the time). Achenes that are full and above a certain size get counted 100% of the time.

In the mini-internship class, Marisol presented her findings to her classmates. See her poster below!

Marisol’s final poster for her mini-internship! Click to enlarge.


Externship Experiences: An Exit Poll

Nina here! Today was Trevor and I’s last day of our externship, which was bittersweet. Although I’m very excited to sleep in past 6:30, I’m definitely going to miss our mentors, Tracy, Lea, and Stuart, as well as all of the volunteers and the whole Botanic Garden community that we briefly got to know during our time here. Trevor and I mostly continued working on our papers, but we also gave short presentations of our projects to the lab to practice presenting. We also had closing interviews regarding our experiences and our thoughts on the externship. Trevor and I have both really enjoyed our time with the Echinacea project, and we hope that you will stay tuned for our final papers, which will be coming very soon (either this weekend or on Monday).

Thanks for everything,

Nina 🙂

Using Our (Green) Thumbs for Typin’

Trevor here! And for, sadly, what is the last time. I can’t believe time flew by so fast! Alas, Nina and I began writing our research papers as we near the end of our externship experience. Our  “mini-papers” will be available on the flog and our broken down into four sections: introduction, methodology, results, and discussion. Make sure to check them out! Nina’s will be about diversity of plant species and its affect on Echinacea seed-set, while mine will concern how the burning of certain sides of a field-site affect the seed-set Liatris and Solidago species. Nina and I also had our “dissertation defense” sessions with Stuart, which both were very enjoyable and left us with lots of answers, but also plenty of questions to consider as we move forward as both students and researchers.

We also got to attend Dr. Kelly Ksiazek Mikenas public dissertation defense discussing her research on green roofs in both Chicago and Germany. In fact, the green roof pictured above is quite famous and is located on top of Chicago’s very own City Hall! Her talk was extremely enlightening and taught me a lot about promoting biodiversity in an urban setting. She defended the use of prairie species on green roofs in cities. She also had a wonderful array of snacks and her resume was quite inspiring and left me lots to think about when it comes to my future.

Thanks for everything Team Echinacea,


Developing achene eye for statistics

Nina here! Today, Trevor and I continued analyzing the data that we collected during the first two weeks of our externship. We both kept working in R, perfecting our papers, and analyzing our results.

Trevor found the confidence intervals for mean seed-set for Liatris and was very excited to see that the two intervals didn’t overlap, strengthening his confidence in his previous findings. Next, he aggregated and merged Solidago data before repeating the analysis he did for Liatris on the Solidago data, but with very different results. It seems as if Solidago was not effected by the burn that occurred in the site (as you can see from the photo on the left), but stay tuned for more updates!

Meanwhile, I looked the diversity function within R’s ‘vegan’ package to use Shannon’s and Simpson’s Diversity Indices on my plant community from each plot. I plotted seed-set, fecundity, and achene count against each index and found the highest correlation when looking at achene count. Next, I calculated species richness as well as the Shannon’s and Simpson’s Diversity Indices including only flowering plants. More updates to follow! After finishing up my actual analysis, I got back to working on my paper (which you can see a preview of on the right) and started thinking about which portions of my results and which graphs I might want to include in my paper.

“R” We There Yet?

Hey! Trevor here! Today, Nina and I continued our work on the R platform, aggregating data and creating figures to be included in our final reports, which (spoiler alert!), will be available on the Flog starting Friday! So if you’re interested in checking out our work, feel free to look at those to get a sense of our projects.

Nina calculated the percent of invasive species for each of her plots using data from the United States Department of Agriculture (USDA) website. She used these percentages to create various graphs to attempt to predict the success of echinacea based on the number of invasive species. However, she found that this was not an effective way of predicting echinacea success. Aside from these quantitative activities, Nina also worked on “beautifying” her graphs, which includes adding R-squared and p values. Look at the red graph to the left, isn’t it beautiful?!


I finished up my work on the top/middle/bottom data and progressed onto the seed counts and cleaning data from the Liatris and Solidago species, which meant that I learned how to merge two different “csv” files together; a challenge, but a challenge I accepted. Afterwards, I worked through some obstacles to effectively calculate the mean seed-set for the East and West sides for the Staffenson site.  East had a higher mean seed set at 0.34, compared to west’s 0.20. In the realm of ecology, that’s a big difference! Currently, I am working on calculating the p value for these findings; it’s a process.

Nina and I hope to finish up working in R tomorrow (we’re almost there!) and begin writing our papers.

Until next time,

Trevor 🙂

“There’s nowhere up but here!”: Externship Week 3

Nina here! To start off the third week of our externship, Trevor and I got back to work in R. We started making our own scripts to try to visualize our data and understand our results. I found a strong correlation between seed-set and fecundity and a correlation between seed-set and achene count (although the R squared value was small). I downloaded a database of plants of Minnesota from the USDA database that included information on whether or not the plants were native, threatened, or endangered as well as their fire tolerance and growth habit. I merged this database with the diversity data collected in the field in order to provide more data about each species. Next, I tried to calculate the percent of invasive and nonnative species for each plot. Meanwhile, Trevor calculated the means for the numbers of achenes in the top, middle, and bottom samples and determined whether or not there was statistical significance between pairwise comparisons of the findings. Next, he looked at confidence intervals for his findings. We both continued our work on our papers, writing up our methods sections specifically. As Trevor meant to say, there’s nowhere to go from here but up!

Friday: A Day of Holiday

Trevor here! Today, Nina and I finished our work in “R” and diverged paths. I recounted Solidago to ensure that my initial counts were accurate. Nina began aggregating her data in “R” and created five extremely beautiful and effective graphs to display her data efficiently. Then, we progressed to the lab meeting, stuffing our faces with salami, cheese, and crackers as we learned about different management methods used with western prairie fringe orchids in Northwest Minnesota. Whilst attempting to answer questions, it seems as if these meetings yield more questions then answers as the intense discussion amongst fellow researchers brings up perspectives not previously contemplated, providing opportunities for future research. Stuffing our faces at lab meeting progressed into stuffing our faces at the holiday party as we joined everyone in the festivities. Highlights include chocolate covered strawberries, mac and cheese, and meatballs. Nina and I, on the recommendation of Leslie and Marty, then proceeded to walk off our meal with a stroll through the garden, visiting the amazing green houses in the Regenstein center. It’s super surreal to be inside a tropical-esque greenhouse in the middle of December! Ending our Friday, we began researching and writing the background information for our future research papers and are excited to continue our writing next week.