I adjusted different models considering the response variable (massaseca) as a function of (tempo) for each treatment level (teor) using the ggplot2 function combined with t