Osmoregulation of arginine vasopressin (AVP) is altered in diabetic ketoacidosis (DKA). With hyperglycemia, the AVP-plasma sodium (PNa) curve is displaced to the left, whereas the AVP-osmolality (Posm) curve is displaced to the right. The shift in the Na curve is explained by either resetting of the Na set point or by glucose acting as a nonpermeable solute, substituting for Na. Conversely, putative unmeasured solutes that, like urea, fail to affect AVP have been postulated to account for the right shift in the AVP-Posm curve. Therefore the respective roles of Posm = sigma [Xi] and plasma tonicity (Pton = sigma [sigmaiXi]), i.e., the sum of concentrations of all solutes [Xi] corrected (Pton) or not (Posm) for their relative cell permeability (sigma i), were studied in DKA. Indeed, Posm = sigma [Xi] exceeds Pton = sigma [sigma iXi] in DKA, since sigma i less than 1 for glucose. Potential determinants of AVP release (Posm, Pton, and PNa) were monitored in 7 patients with DKA. Conventional correlation analysis and two-dimensional (2D) graphs reproduced the paradox of an opposite shift in PNa and Posm set points for AVP release. However, by using the concept of tonicity instead of osmolality, 3D plots instead of 2D graphs, and multiple regressions instead of correlations, the AVP-PNa and AVP-Pton curves did not appear displaced. The concept of tonicity resolved the paradox of both osmolality and Na thresholds reset in opposite directions. Indeed, in states where a solute like glucose (with sigma less than 1) contributes substantially to plasma osmolality, Posm measured in vitro by the osmometer greatly exceeds Pton perceived in vivo by the osmoreceptor.