Under the null hypothesis of the independence of hypertension and smoking status, we have the following expected table (rounding off the expected frequencies)

\(\begin{array}{|c|c|} \hline Tension\ level & Non-smoker & Moderate\ smoker & Heavy\ smoker \\ \hline Hypertension & 33 & 31 & 22 \\ \hline No\ hypertension & 37 & 34 & 24 \\ \hline \end{array}\)

This gives the value of T to be 16.486, which, under the null, follows an asymptotic chi-squared distribution, with 2 degrees of freedom. This yields a p value of 0.00026, and hence we reject the null hypothesis.