The habitat categories 31 and 34 most 
accurately represent the conditions of the 
woodlands surveyed by Woolhouse. The total 
area of these habitats can thus be used to 
predict bird species numbers. However, these 
conclusions can only be accepted tentatively. 
This apparent relationship may have occurred 
purely by chance. Unfortunately field data 
only exists for one year therefore the 
relationship cannot be retested. A backup 
test was conducted. If the relationship 
between the total habitat area of categories 
31 and 34 and the number of breeding species 
was real, one would expect all the bird 
species which breed to have either 31 or 34 
as one of their habitat requirements. It was 
found that 20 of the 27 species known to 
breed in Saltswells wood had habitats 31 or 
34 listed as one of their preferred habitat 
types. This leaves the breeding of 7 species 
unexplained. Two conflicting conclusions can 
be drawn. The apparent relationship between 
the actual number of breeding bird species 
and the number estimated using the Woolhouse 
equation has occurred by chance or the 
relationship is real but the habitat patterns 
allocated to each bird species are 
inaccurate. The results of this analysis 
remain therefore inconclusive and indicate 
that further work needs to be conducted. 
Woolhouse, in the same paper also suggests 
that species abundance can be predicted from 
the area of woodland. 
In I = 0.679 In A + 3.082 
Where In I is the natural log of species 
abundance. 
Using the total habitat area of categories 31 
and 34 the expected number of individuals is 
174.12. However, according to the field data 
289 individual birds were recorded. This 
large discrepancy suggests that this equation 
may not be used with the area totals of 31 
and 34 identified from the air photographs 
for urban woodlands. The field data 
collected by Woolhouse was obtained in rural 
woodlands. Various studies about bird 
populations in urban environments for 
example, Emlen 1974, Tomialojc and Profus, 
1977 suggest that although species numbers 
are lower in urban environments than rural, 
actual bird densities are often higher in 
urban areas. As Saltswells wood is an urban 
wood this may be the cause of the 
discrepancy. 
There is not enough evidence available to 
indicate whether the equation put forward in 
the literature concerning the relationship 
between area and bird species numbers can be 
used successfully with air photographs. The 
results of the analysis are inconclusive. 
However this does not necessarily suggest 
that it is impossible to predict the bird 
carrying capacity of a wood from habitat area 
data extracted from air photographs. 
Further analysis was conducted by 
calculating the total amount of suitable 
habitat available in the reserve for each 
likely bird species. 
The photographic area data suggested that the 
species which were identified as breeding in 
the field survey were those which had the 
largest area of suitable habitat available in 
the reserve. Two statistical tests, Chi^ and 
Mann Whitney U test were conducted to 
investigate whether the available habitat 
totals for breeding and non-breeding species 
were from the same population or whether 
there was a significant difference between 
Table 2. Species habitat area 
Bird 
Species 
Suitable Habitat 
categories 
Total 
area 
habitat 
(ha) 
Black bird 
34/43/44/45 
16.95 
Woodwarbler 
31 
14.43 
Willow 
Warbler 
44/45 
1.02 
the two groups. The rejection level in both 
instances was 0.05. 
The Chi^ test suggested there was no 
difference between the area totals but the 
Mann Whitney U test indicated that there was 
a significant difference, that is, the bird 
species which are found to breed in the 
reserve are those with the largest amount of 
habitat available to them. However this test 
does not indicate the magnitude of the 
difference, nor does it provide any 
indication of a critical value of available 
area above which bird species are likly to 
breed. The parameters needed for prediction 
have not been identified. 
Further tests were thus felt necessary to 
discover whether any other differences could 
be identified from air photographs between 
the habitats of breeding and non-breeding 
species. It was believed the level of 
dispersion of suitable habitat types may have 
an influence on whether a bird species 
breeds. The previous tests have looked at 
the relationship between the numbers of 
breeding birds and the total amount of 
suitable habitat. However in reality, in the 
reserve, the habitat types are fragmented 
into a number of small units rather than 
individual large blocks. 
This fragmentation may result in suitable 
habitat blocks being below the territorial 
requirements of an individual species so 
being of little use. It was therefore 
suggested that the bird species with the 
least divided habitat will be more likely to 
breed. They will have more chance of finding 
a suitable sized block of the required 
habitat type. 
It is easy to identify the number of units 
of each habitat type from the air 
photographs. Mann Whitney U tests were 
carried out on data from breeding and 
non-breeding birds. The tests looked at: 
a) whether any significant difference 
existed between the numbers of units of 
suitable habitat for breeding and 
non-breeding bird species, 
b) whether any significant difference 
existed between the amount of area per unit 
and 
c) whether any significant difference 
existed between the average number of units 
per habitat type for breeding and 
non-breeding species. 
In each instance the rejection level was 
set at 0.05. In each case no significant 
difference was found. No conclusison could 
therefore be drawn about the affects of 
habitat unit dispersion and the likelihood of 
birds breeding in the urban woodland. 
6 CONCLUSION 
There is a great deal of discussion in the 
literature concerning the usefulness of 
dif ferei 
numbers 
diversit 
any agrt 
using a: 
tests ce 
identif 
number 
could b 
the typ 
their di 
the re 
suggest: 
The nex 
look at 
identify 
populat 
many of 
air phol 
combine 
7 REFEP 
Aronoff 
as ai 
Photc 
Sensi 
Arrheni 
Jourr 
Emlen J 
Tucsc 
regul 
Fosberg 
class 
purpc 
IBP a 
Ginevan 
accui 
Engii 
10:12 
Lancast' 
comrru 
habit 
57,12 
Moore N 
numbe 
Biolc 
Nature 
Natui 
state 
Skaley i 
with 
Unive 
Tomialo; 
anal^ 
two 
Quer< 
Ornit 
West Mi 
Natu: 
Count 
Repor 
Witherb 
Handt 
and G 
Woolhou 
pract 
appli 
wood 
27,4 :