correlation differ markedly from that found in the 1950 survey, and even here the relationship showed up as statistically significant at the a = 0.05 significance level. Distance separating the sample cities from their neighbours of the 'same size' (Hypothesis I) was also compared with distances separating sample cities from their neighbours of the 'same or greater size'(Hypothals II).
Figure 4.194 shows the correlation values obtained on the first test against those obtained on the second. While for both sets of values the coefficients are statistically significant (a = 0.05 significance level), the closer relationship is always obtained for the second test. It is dear for Iowa that town size is associated with spacing, and that this relationship holds in a hierarchical sense, since the sample towns are more closely associated with their larger neighbours. Whether the closing gap shown by the two sets of relationships represents a trend to a more highly integrated relationship between the settlements remains to be seen.