R. J. BING AND K. D. HELLBERG 
809 
We have repeatedly stressed the difficulties 
inherent in the method aside from those intro- 
duced by a continuously moving objective. There 
is the uncertainty that the microvascular pat- 
tern in the left atrium of the cat, which is sap- 
plied by branches of the left circumflex coro- 
nary artery, is identical to that in other portions 
of the heart muscle. A further difficulty is that 
it is necessary to apply a certain degree of pres- 
sure against the muscle in order to prevent 
flooding of blood between the hollow glass tube 
and the inner surface of the atrial wall. How- 
ever, in a series of experiments, we determined 
that the tension was not sufficient to prevent 
free movement of red cells in the capillaries. A 
third criticism which has been raised is that 
there may be heat conduction from the xenon 
arc to the wall of the left atrium. We have 
shown, however, that the temperature at the 
! atrial end of the conducting system differs by 
less than 2° C from that of the blood. Red cell 
velocity was measured by frame-to-frame anal- 
ysis of the movements of red cells by means of 
a film analyzer. We calculated average data of 
I from between 6 to 40 observations per experi- 
I mental series. Determinations could only be 
made on capillaries which remained in focus 
throughout the cardiac cycle (usually from 3 
to 6 capillaries). The data were then averaged 
and statistically treated. 
In a series of experiments, we have observed 
the microcirculation of the potassium-chloride- 
arrested left atrium of the cat.^ Perfusion was 
carried out by means of a Sigma pump using 
heparinized cat's blood. 
RESULTS AND DISCUSSION 
The pattern of the coronary microcirculation 
is illustrated in Figure 2. Our main efforts were 
directed toward an investigation of the rela- 
tionship of flow from one capillary to the other. 
In 1919, Krogh, in collaboration with Erlang, 
derived a mathematical solution for steady-state 
radial diffusion from a cylindrical capillary to 
the tissue cylinder around it." The Krogh- 
Erlang equation originated from the assump- 
tion that the capillary is infinitely long and 
I straight. It is still used as a fair estimate for 
1 an approximate description of the oxygen sup- 
ply from the capillaries into the surrounding 
tissues. It therefore assumes a cylindrical sym- 
metry of capillary arrangements with parallel 
unidirectional flow of blood in the capillaries. 
On the basis of our studies, these assumptions 
do not appear to be valid. This is because of 
the presence of countercurrent flow in the capil- 
laries of the heart and because of asymmetric 
capillary arrangement. 
In the perfused heart, as well as in the nor- 
mally beating organ in situ, there occurs coun- 
tercurrent flow in adjacent capillaries. This is 
the case primarily in capillaries with intercon- 
necting loops. Similar findings have been ob- 
served by Grote and Huhmann^^ ^nd by Huh- 
mann and Niesel in the perfused rat heart. 
In addition, we found that capillaries in over- 
laying muscle layers form adjoining sheets. 
Therefore, one can speak of asymmetric capil- 
lary arrangement with occasional countercur- 
rents. This arrangement provides a favorable 
oxygen distribution to the cardiac muscle. 
A further observation which is of significance 
Figure 2. 
