Statistics
The data were analyzed using XLstat 2019: Data Analysis and Statistical Solution for Microsoft Excel (Addinsoft, Paris, France). One-way analysis of variance (2016 data) or linear mixed-effects models (2017 data) were performed on the data for amino acids, polyamines, inorganic ions, soluble sugars, chlorophylla +b , total soluble protein, and leaf nitrogen followed by Tukey’s HSD post hoc test to determine if differences existed among treatment groups on a given sampling date. The fixed factor wastreatment for all sampling dates and the random factor wasgreenhouse room for only the 2017 sampling dates. This statistical approach was used instead of analyzing a 2x2 factorial design to avoid potential issues of pseudoreplication in the first year when there was only one greenhouse room per temperature treatment.