To the Editor: Drs Jenkins and Campbell recall a basic equation of glucose homeostasis (equation 1 in Jenkins and Campbell [1]) that is indeed useful for understanding the interrelationships between beta cell function, insulin sensitivity and glucose tolerance. This equation illustrates well their critique of the disposition index, on which we have also expressed some reservations [2, 3]. However, we think that their analysis may miss a central point, i.e. the nature and physiological meaning of these relationships. They say that the disposition index ‘theory was originally developed to explain the hyperbolic relationship between insulin secretion and insulin sensitivity, which within our model is an intrinsic property resulting from the underlying physiology requiring no additional mechanistic explanation’ [1]. It seems to us that this view ignores the cause–effect relationship. It is the existence of a physiological inverse relationship between beta cell function (β) and insulin sensitivity (Si) that makes glucose (G) relatively constant, as G is determined by β and Si (assuming that glucose appearance, Rin in Eq. 1 of Jenkins and Campbell [1], is of minor importance). If, conversely, β and Si were unrelated, G would not be equally constant, but obviously the lack of the β–Si relationship would be the primary fact and not a consequence of a relationship masked by large variations in G. These arguments apply to Fig. 1c of our article [4]: we think that the lack of a significant relationship between ln(β) and ln(Si), when the glucose range (ln(G)) is not too narrow, is not due to the existence of an underlying β–Si relationship masked by ln(G) variations. On the contrary, it is the lack of relationship between ln(β) and ln(Si) that makes ln(G) so strongly dependent on ln(β). In other words, equation 1 of Jenkins and Campbell [1] is a constraint that can be read: ‘if Si is inversely related to β then one can expect that glucose is relatively constant; otherwise, glucose will vary in relation to β and Si’. Thus, it is the existence or non-existence of a physiological relationship between insulin sensitivity and beta cell function that will determine the glucose levels. The correlation between insulin sensitivity and beta cell function is subject to direct testing, as always, keeping in mind the statistical characteristics of the indices used. In our study [4], the expected estimation errors associated with the various beta cell function parameters are not very different [5], yet in 1,123 individuals with normal glucose tolerance (NGT) glucose sensitivity was not correlated with insulin sensitivity while several other parameters were (including, as expected, the model surrogate of the acute A. Mari (*) ISIB-CNR, Corso Stati Uniti 4, 35127 Padova, Italy e-mail: andrea.mari@isib.cnr.it