New primary data on the behaviour of Canadian stock prices since the turn of the 19th century are combined with the existing Toronto Stock Exchange Stock Price Return Index to provide a stock price index that spans the full length of the 20th century. Although long period returns have been available for several countries, to date no such data have been published for Canadian markets notwithstanding the fact that equity trading was quite active during this period, especially in Montreal. With these data we are in a position to address for the first time a number of anomalies associated with long horizon returns and irregularities that have been reported for other countries, such as the Gibson Paradox and the January effect. Results only partially support the Fisher hypothesis, but suggest that both the Gibson Paradox and the January effect no longer hold. A December effect appears to have replaced the January effect in the last decade of the 20th century.