16 
Cunninghamia 
Vol. 1 (1): 1981 
Although no attempt was made to record all species present in the area, an 
idea of the total flora to be expected may be estimated from a comparison with the 
adjacent Dharug National Park, an area of similar size. Here Matthew (1973), in 
three years of botanical exploration, recorded 514 native plant species. In Man¬ 
grove Creek only 196 species have been listed, including six exotic species, and it 
is reasonable to believe that a list of similar size to Matthew's could be compiled. 
Classification of communities 
Comparisons between the plant communities described for Mangrove Creek 
and relevant communities described by other workers are shown in Table 4. Pidgeon 
(1937) described two Associations for the area, while Specht et al. (1974) probably 
largely based on Pidgeon, described three. Pidgeon (1940), however, suggested 
that her Associations could be subdivided into “forest types”, each of which would 
have a similar floristic composition of dominants and develop in essentially similar 
habitats. She did not, however, suggest any “forest types” for the Mangrove Creek 
area. The four communities recognized in the present survey are essentially “forest 
types” though each probably contains more habitat variation than she would have 
intended. Baur’s (1965) “Forest Types” are similarly defined and types can be 
correlated with each of the communities recognized for Mangrove Creek. 
However, although this level of classification was satisfactory for descriptive 
work, it was unsatisfactory for mapping. This is due mainly to the particular 
topography of the dissected sandstone plateaux, features of which arc narrow, linear 
units along creek and gullies, and marked aspect differences, making up a recurring 
mosaic of communities often with intergradation between them. For mapping, 
medium-scale colour aerial photography and a 1:25 000 scale topographic base map 
were used, but only structural vegetation units could be consistently recognized 
(Table 1) and even then the differentiation between some of these (tall open-forest 
and open-forest) was unsatisfactory. It appears that although forest types are 
recognizable on Hawkesbury Sandstone, it is very difficult to map units within 
structural formations, except where these can be related to topographic or geological 
factors, which are otherwise evident in aerial photography. Similar results are in¬ 
dicated by J. S. Benson and Fallding (1979) who mapped Hawkesbury Sandstone 
areas in Brisbane Water National Park, about 20 km south-east of Mangrove Creek. 
Dealing with a more varied area, both floristically and structurally, ranging from 
closed-forest to sedgeland, they mapped (at scale 1:25 000) 18 communities, almost 
all recognized on structure with some subdivisions related to topographic position. 
Forest types on Hawkesbury Sandstone are therefore probably best treated as 
components of a vegetation gradient associated with slope, soil depth, aspect etc. 
Austin (1978), in dealing with the sclcrophyll forest communities of the South Coast 
of New South Wales, successfully grouped his forest communities into “types” on 
the basis of toposequences and altitudinal range. Using a similar approach at 
Mangrove Creek one would also take into consideration the major geological units 
present, as these are associated with important floristic differences. Thus a Eucalyp¬ 
tus tereticornis-E. deanei-Angophora floribunda toposequence type could be recog¬ 
nized on Narrabeen Group shales and sandstones, and a £. eximia-E. punctata- 
Angophora costata type on Hawkesbury Sandstone. Such types may need to be 
modified or extended when applied to areas with greater topographic variation than 
Mangrove Creek, for instance, on Hawkesbury Sandstone in the vicinity of the 
Colo River a possible toposequence-type would be E. eximia-E. punctata-Angophora 
costata-E. saligna. 
Once toposequence types have been established, the relationships between 
vegetation structure, floristic composition and various gradients within these types 
can be studied in greater detail. 
