Regardless of the recent trends regarding the design of multi-objective evolutionary algorithms (MOEAs), non-dominated sorting remains as an important mechanism to classify solutions to a multi-objective optimization problem. Non-dominated sorting is defined for a set of points in M-dimensional space (where M is the number of objectives of a multi-objective problem). These points are the objective vectors associated with the solutions of the optimization problem. Many approaches have been proposed in the last two decades to produce more computationally efficient non-dominated sorting procedures. A recent approach is Merge Non-Dominated Sorting (MNDS) which uses the concept of dominance set of points and exploits the efficient implementation of bitset in the Java programming language. In the original paper proposing MNDS, it was claimed that for the best case to happen, there should be no dominance among the points, i.e., all points should be in a single front. However, in this paper, we show that, in this particular scenario, the worst-case can also occur. So, when there is no dominance among the points, i.e., all the points are in the same front, the best and the worst cases can occur. We also provide here a detailed complexity analysis of this algorithm.