For this experiment we chose white spruce as the primary species due to its abundant use in reforestation and vast amount of knowledge around its growth. The forb species chosen are fireweed and showy aster. They are both native species that are first colonizers of disturbed areas. Two different sizes of plug were chosen. The larger plugs were deemed necessary due to the increased number of roots present with the forb inclusion. The control stocktype was grown in the industry standard plug.
Besides the two forb species we had two independent variables (Figure 1), sow date (6, 8, 10 or 12 weeks) and plug size (512A or 615A). The sow date refers to the time between sowing the spruce and the forb. The head start was deemed necessary due to the rapid growth of the forb compared to the spruce. The plug sizes are two larger sizes (512 refers to volume, 512ml) to allow room for the roots. The control was grown in the industry standard 412A plug. This led to a total of 16 treatments plus the control (17 stocktypes).
Each site was laid out in a series of 7 lines (Figure 2). The lines are 2m apart, and each line has all 17 stocktypes in a random order. Each stocktype in each line consists of 3 replicates. All trees are planted 2m apart in each line. Measurement data was gathered on the entire planted population. Every tree was individually measured for height, leader growth and root collar diameter. A 0.5m by 0.5m quadrant was placed over every tree and within that quadrant fireweed and aster max height was measured. Additionally in that quadrant fireweed, aster, native herbaceous species, non-native herbaceous species, and grass species cover were estimated on a percent basis.
Figure 1. Flow chart of the breakdown of the 16 different treatments.
Figure 2. Diagram of the layout of the 17 stocktypes (16 treatments + control) on every site.
In this study we focused on how the hitchhiker species compare to the control group, and was not concerned about how they compare to each other. For this reason it was determined that Dunnett's test for multiple comparisons was the best method for assessing both stocktype viability and cover of non-native herbaceous species. This test allowed us to compare each treatment to the control and accounts for the increased probability of type one errors. The benefit of Dunnett's test is it doesn't result in an unnecessarily wide confidence interval like Tukey's and Scheffe's methods.
For assessing the cover of the target forbs, individual two-sided t-test's were used for simplicity. Only the stocktypes that showed to be viable based on the Dunnett's tests were compared to the control. This method was chosen as it allowed for a simple comparison of forb cover of the viable stocktypes to the control. We are not concerned about interaction effects here as the only conclusions that will be made or whether some stocktypes are viable, and not which is the best stocktype overall. All statistical analysis and output was done using R Software (R Core Team, 2016).