In 11 normal and in 15 dehydrated unanesthetized mongrel dogs, shock was produced by graded hemorrhage. Hemoconc. occurred in 18% of the normal and in 53% of the dehydrated animals. This hemoconc. resulted from the net loss of additional plasma fluid and protein as shock developed. Typical gross and histologic findings of shock were iound in a number of animals. The hemodiluting non-dehydrated animals tolerated an avg. blood loss of only 49%; of their initial blood vol.; whereas only 43% could be withdrawn from the others. The hemoconcentrating dehydrated animals hemodiluted during and after hemorrhage less than the others; this could explain their inability to tolerate as much bleeding before going into shock and could lessen the masking by earlier hemodilu-tion of the subsequent hemoconc. It is suggested that the conflicting reports as to the occurrence of hemoconc. after hemorrhage may be related to the state of hydration of the exptl. animal. There are no definite grounds for differentiating between hemorrhagic shock and shock from other causes. In modern warfare dehydration is not uncommon, and severe hemorrhage may produce shock with hemoconc. In such a case, the clinician who has been taught that only hemodilution can follow hemorrhage may falsely assume that the patient has not lost much blood. The opposite may be the case.