At below replacement level, fertility changes are subtle and complex; and distinguishing statistically significant trends from random shifts is becoming a relevant issue. The probabilistic fertility table describes the uncertainty of the childbearing process, and provides a significance test for the annual changes of various fertility measures, which is essential for distinguishing between a statistically significant change from a random fluctuation. This paper provides an analytical model for the total fertility of the probabilistic fertility table, and extends the significance test to period trends that include multiple annual changes. The extended significance test indicates that complex annual changes could accumulate to become a significant trend. Applying the analytical model and extended test to the total fertility of Canada, it indicates that the 2000–11 upward trend is statistically significant and, therefore, supports recently projected future increases of total fertility