256 
PULMONARY MODELS 
Figure 1. — Examples of capillary tracings on the surface of the dog lung in vivo. Thin lines are alveolar walls 
that abut with pleural surface. Heavy lines are perfused capillaries. Note greater perfusion on bottom of 
lung compared to top of lung under all conditions. Hypoxia causes increased capillary perfusion both on top 
and bottom of lung. Each square in grid (upper left) is 50fi X 50/i. 
does indeed occur and that more recruitment 
takes place at the top of the lung than on the 
bottom. What is less certain is the reliability of 
our grid count method for estimating the mag- 
nitude of the recruitment response. The method 
is not objective because the observer has to 
judge v^^here the capillaries are located and 
whether they are perfused. In many cases the 
perfused capillaries are obvious, but often 
subjective judgment is required. The reason 
these data must be considered preliminary is 
that only one observer has made a grid count 
analysis of the films. Undoubtably the trends 
indicating recruitment will be present in fur- 
ther analyses, but the actual numbers will al- 
most certainly change. Careful statistical analy- 
sis will be required to determine whether the 
magnitude of recruitment as indicated by the 
grid count technique is trustworthy. 
Changes in capillary diameter, erythrocyte 
velocity, capillary hematocrit, and red cell ori- 
entation are unmeasured. Subjectively, there do 
seem to be changes in these parameters during 
hypoxia. However, recruitment seems to be the 
dominant form of change in capillary volume 
under these conditions, especially at the top of 
the lung. 
Given these reservations, some discussion of 
the data is possible in an attempt to determine 
the mechanism that causes recruitment with 
hypoxia. There are three apparent mechanisms 
either acting separately or in concert that could 
