Observation Of College Seeds

In September, we went the Berry College chestnut tree orchard to collect seeds from many of the trees present in the orchard. Around 900 seeds were collected in total, approximately 25 from each seed lot. Using tree pruners, clusters of burs were cut down so that the seeds could be harvested. The clusters we cut off branches were partially open. Students pried open the burs, careful not to damage the style on the seeds, and harvested the good seeds. Some burs contained seeds that had not matured properly; these seeds were not used. The viable seeds were then placed in a Ziploc bag with the correct cross code labeled on the front of the bag. The seeds collected were from the following trees: AM, CH, F1, BC3s, and BC4s. Once the seeds were collected

they were sealed in bags with moist, sterile sphagnum at +/- four degrees Celsius until measurements were to be taken.
The seeds were then disbursed to students in a lab setting in the previously mentioned Ziploc bags, and they were placed on ice as they had to be kept cold. Each student measured approximately 25 seeds each. Using Vernier calipers, we measure the length, width, and depth of each nut. Then we measured the width and height of the hilium scars. To measure the length of the longest “arm” on the hilium scars, a dissection scope was required as the arms on the hilium are small. The hairiness of each nut was recorded as a percentage of 0-100%, zero meaning no hair and 100 meaning the nut was completely covered. Each seed was determined to either be a side or center nut and recorded along with the other measurements. Lastly the length of the stigma/style was measured. All measurements, except hairiness, were recorded to 0.01mm.
The data collected from the measurements were then used to make further calculations used for comparison.

The seed volume from average radius was found using the equation VOL = (4/3)*3.1416* ((L/2+W/2+D/2)/3)3. Surface area was calculated using an approximate formula, SA~ 4pi [(a^p b^p + a^p c^p + b^p c^p )/3]^{1\p}, where p=1.6075, and a, b, c are the three radii (a>=b>=c), made by Knud Thomsen. Hilium area was found by dividing the hilium width and the hilium height by two then multiplying by each other and 3.1416. By dividing the hilium width by the surface area, we found the hilium area to surface area ratio. Hilium width to seed width ration was calculated by dividing the hilium width by the seed width. Arm length was divided by hilium width to get the arm length to hilium width ratio, and finally stigma/style length to seed length ratio was calculated by dividing style length by seed length.
To compare means of each above variable, two things were used, non-parametric Kruskal-Wallis tests, and Dunn’s All Pairwise Comparisons. These compared each type of cross to one another. This allowed us to determine if any minor difference, or none, were present in the pure American chestnut and the hybrid backcrosses. As well, we were able to tell if any big difference came up between those two lines and the Chinese chestnut

line.
Then to compare morphological characteristics for all, we used principal components analysis. Each new variable, principal components, created was a combination of all the other variables. A principal component was created for each line. The variables were compared, and a percentage was calculated for the variation from the original variables that was incorporated into the new variables. The new variables were plotted, and checked to see if the American and backcross lines were close and that they were separate from the Chinese line. To compare each line to another Bonferroni All Pairwise Comparisons was used.