We present empirical evidence of isospin dependence in the compound nuclear fission cross-sections and fission widths, which suggests that the compound nucleus (CN) possibly retains the memory of the isospin when it is formed. We examine the idea, first proposed by Yadrovsky [1], for three pairs of reactions where experimental data of fission cross section at various excitation energies are available. One of the pairs of reactions is the same as used by Yadrovsky i.e. 209 Bi(p, f) and 206 Pb(α, f) leading to the CN 210 Po but with an improved experimental data set. The other two pairs of reaction sets are, 185 Re(p, f) and 182 W(α, f) leading to the CN 186 Os and, 205 Tl(p, f) and 202 Hg(α, f) leading to the CN 206 Pb. An observable difference between the fission branching ratios in two different isospin states suggests that the CN seems to remember its isospin at the point of formation. This possibility is further supported by another method, where additional empirical evidence for four CN, viz. 210 Po, 209 Bi, 207 Bi, and 198 Hg, is obtained from the experimental data in Zhukova et al. [2]. Further, the data also suggest a possible new signature of the weakening of CN process and gradual transition to non-compound processes as the energy rises. Fresh experimental efforts as proposed, are required to confirm these findings.