American cities, Consistent negative 
density averages and owner-occupancy, 
occupational status groups. In Birmingham, the description of resi- 
dential subareas in terms of 
essentially defined the ecologi 
these correlations demonstrat 
and urban social Structure, 
to have greatest predictive 
correlations were found between 
income, and proportions of high 
"high" and "out dwelling-unit density 
cal pattern of racial segregation. While 
€ associations between dwelling-unit density 
this particular data category, was found 
value with respect to population size and 
density rankings of residential neighborhoods, 
As might be expected, for all cities studied, the correlations 
between average number of dwellings 
average number of persons 
«95 and above. 
tion estimates for residenti 
For example, with reference 
Birmingham subareas accor 
per block (by census tracts) and 
per block were consistently on the order of 
This fact provides a basis for deriving certain popula- 
al neighborhoods studied photographically, 
to Table 2 above, the ranks of the seventeen 
ding to photo data on dwelling-unit density 
correlate .95 with their ranks on population density, 
In another evaluation of thig item, forty-three Rochester, New 
York census tracts we 
re grouped into five dwelling-unit density classes 
ranging from "very low!" to "very high", the accuracy tests having indicated 
the validity of similar i 
tation. The resident popul 
‘per block was computed as shown below, 
An analysis of varianc 
nformation collected by photographic interpre- 
ation density in average number of persons 
  
Dwelling-Unit Average Number of 
Density Persons per Block 
Very Low 89 
Low 123 
Medium 153 
High 175 
Very High 246 
e test showed that these sharp differences in 
resident population density are highly significant statistically. These 
results thus indicate that the photo i 
neighborhoods according to these dwell 
à meaningful description of population 
complex, 
Multi-Variate Analy:-g of 
The several ecological analyses b 
strated that the various physical stru 
the four photo data ite 
of the city, 
nterpreter!s classification of 
ing-unit density categories provides 
density patterns within the urban 
Urban Structures 
riefly described above demon- 
ctural categories derived from 
Ms are useful for defining the social topography 
However, generally speaking, any one îtem, as a single 
1)