We review evidence supporting the conclusion that renal dysfunction un
derlies the development of all forms of hypertension in humans and exp
erimental animals. Indexes of global renal function are generally norm
al in the early stages of most genetic forms of hypertension, but rena
l function is clearly impaired in long-established hypertension. Studi
es in our laboratory over the past decade summarized below have establ
ished that the renal medulla plays an important role in sodium and wat
er homeostasis and in the long-term control of arterial pressure. Deve
lopment of implanted optical fibers for measurement of cortical and me
dullary blood flows with laser-Doppler flowmetry and techniques for de
livery of vasoactive compounds into the medullary interstitial space e
nabled us to examine determinants of medullary flow (nitric oxide, atr
ial natriuretic peptides, kinins, eicosanoids, vasopressin, renal symp
athetic nerves, etc). We have shown in spontaneously hypertensive rats
that the initial changes of renal function begin as a reduction of me
dullary blood flow in the absence of changes of cortical flow. Long-te
rm medullary interstitial infusion of captopril, which preferentially
increased medullary blood flow, resulted in a lowering of arterial pre
ssure. In normal Sprague-Dawley rats, selective reduction of medullary
flow with medullary interstitial or intravenous infusion of small amo
unts of N-G-nitro-L-arginine methyl ester resulted in hypertension. Th
ese and other studies we review show that although blood flow to the i
nner renal medulla comprises less than 1% of the total renal blood flo
w, changes in flow to this region can have a major effect on sodium an
d water homeostasis and on the long-term control of arterial blood pre
ssure.