The use of static human body dimensions to assess the human accessibility is an essential part of an ergonomic approach in user-centered design. Assessments of reach capability are commonly performed by exercising external anthropometry of human body parts, which may be found in anthropometric databases, to numerically define the reach area of an intended user population. The result is a reach envelope determined entirely by the segment lengths, without taking into account external variables, as the nature of the task or the physical capacities of the subject, which may influence the results. Considering the body as a simple assembly of static parts of different anthropometry is limiting. In this paper, the limit of validity of this approach is assessed by comparing the reach envelopes obtained by this method to those obtained with a simple two-dimensional experimental reaching task of a panel of subjects. Forty subjects experimentally evaluated the reach, first with the body constrained and second unconstrained. Results were recorded and compared with those obtained numerically with a model, based on their own anthropometric characteristics, previously measured. A statistical study of the results allowed the definition of the shape of a confidence bound containing the real reach envelope. The results indicated important differences between the experiment and the numerical evaluation of the reach envelope.